feat: per-project LLM models, ACP session pool, smart pagination tests

- ACP session pool with per-model subprocesses and 8h idle eviction
- Per-project LLM config: local override → mcpd recommendation → global default
- Model override support in ResponsePaginator
- /llm/models endpoint + available models in mcpctl status
- Remove --llm-provider/--llm-model from create project (use edit/apply)
- 8 new smart pagination integration tests (e2e flow)
- 260 mcplocal tests, 330 CLI tests passing

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

completions/mcpctl.bash

@@ -2,10 +2,10 @@ _mcpctl() {
   local cur prev words cword
   _init_completion || return
 
-  local commands="status login logout config get describe delete logs create edit apply backup restore mcp help"
+  local commands="status login logout config get describe delete logs create edit apply backup restore mcp approve help"
   local project_commands="attach-server detach-server get describe delete logs create edit help"
   local global_opts="-v --version --daemon-url --direct --project -h --help"
-  local resources="servers instances secrets templates projects users groups rbac"
+  local resources="servers instances secrets templates projects users groups rbac prompts promptrequests"
 
   # Check if --project was given
   local has_project=false
@@ -78,7 +78,7 @@ _mcpctl() {
   case "$subcmd" in
     config)
       if [[ $((cword - subcmd_pos)) -eq 1 ]]; then
-        COMPREPLY=($(compgen -W "view set path reset claude impersonate help" -- "$cur"))
+        COMPREPLY=($(compgen -W "view set path reset claude claude-generate setup impersonate help" -- "$cur"))
       fi
       return ;;
     status)
@@ -114,7 +114,7 @@ _mcpctl() {
       return ;;
     create)
       if [[ $((cword - subcmd_pos)) -eq 1 ]]; then
-        COMPREPLY=($(compgen -W "server secret project user group rbac help" -- "$cur"))
+        COMPREPLY=($(compgen -W "server secret project user group rbac prompt promptrequest help" -- "$cur"))
       fi
       return ;;
     apply)
@@ -150,6 +150,15 @@ _mcpctl() {
       fi
       COMPREPLY=($(compgen -W "$names" -- "$cur"))
       return ;;
+    approve)
+      if [[ -z "$resource_type" ]]; then
+        COMPREPLY=($(compgen -W "promptrequest" -- "$cur"))
+      else
+        local names
+        names=$(_mcpctl_resource_names "$resource_type")
+        COMPREPLY=($(compgen -W "$names" -- "$cur"))
+      fi
+      return ;;
     help)
       COMPREPLY=($(compgen -W "$commands" -- "$cur"))
       return ;;

completions/mcpctl.fish

@@ -3,7 +3,7 @@
 # Erase any stale completions from previous versions
 complete -c mcpctl -e
 
-set -l commands status login logout config get describe delete logs create edit apply backup restore mcp help
+set -l commands status login logout config get describe delete logs create edit apply backup restore mcp approve help
 set -l project_commands attach-server detach-server get describe delete logs create edit help
 
 # Disable file completions by default
@@ -28,7 +28,7 @@ function __mcpctl_has_project
 end
 
 # Helper: check if a resource type has been selected after get/describe/delete/edit
-set -l resources servers instances secrets templates projects users groups rbac
+set -l resources servers instances secrets templates projects users groups rbac prompts promptrequests
 
 function __mcpctl_needs_resource_type
     set -l tokens (commandline -opc)
@@ -36,11 +36,11 @@ function __mcpctl_needs_resource_type
     for tok in $tokens
         if $found_cmd
             # Check if next token after get/describe/delete/edit is a resource type
-            if contains -- $tok servers instances secrets templates projects users groups rbac
+            if contains -- $tok servers instances secrets templates projects users groups rbac prompts promptrequests
                 return 1 # resource type already present
             end
         end
-        if contains -- $tok get describe delete edit
+        if contains -- $tok get describe delete edit approve
             set found_cmd true
         end
     end
@@ -55,12 +55,12 @@ function __mcpctl_get_resource_type
     set -l found_cmd false
     for tok in $tokens
         if $found_cmd
-            if contains -- $tok servers instances secrets templates projects users groups rbac
+            if contains -- $tok servers instances secrets templates projects users groups rbac prompts promptrequests
                 echo $tok
                 return
             end
         end
-        if contains -- $tok get describe delete edit
+        if contains -- $tok get describe delete edit approve
             set found_cmd true
         end
     end
@@ -139,6 +139,7 @@ complete -c mcpctl -n "not __mcpctl_has_project; and not __fish_seen_subcommand_
 complete -c mcpctl -n "not __mcpctl_has_project; and not __fish_seen_subcommand_from $commands" -a apply -d 'Apply configuration from file'
 complete -c mcpctl -n "not __mcpctl_has_project; and not __fish_seen_subcommand_from $commands" -a backup -d 'Backup configuration'
 complete -c mcpctl -n "not __mcpctl_has_project; and not __fish_seen_subcommand_from $commands" -a restore -d 'Restore from backup'
+complete -c mcpctl -n "not __mcpctl_has_project; and not __fish_seen_subcommand_from $commands" -a approve -d 'Approve a prompt request'
 complete -c mcpctl -n "not __mcpctl_has_project; and not __fish_seen_subcommand_from $commands" -a help -d 'Show help'
 
 # Project-scoped commands (with --project)
@@ -157,7 +158,7 @@ complete -c mcpctl -n "__fish_seen_subcommand_from get describe delete; and __mc
 complete -c mcpctl -n "__fish_seen_subcommand_from edit; and __mcpctl_needs_resource_type" -a 'servers projects' -d 'Resource type'
 
 # Resource names — after resource type is selected
-complete -c mcpctl -n "__fish_seen_subcommand_from get describe delete edit; and not __mcpctl_needs_resource_type" -a '(__mcpctl_resource_names)' -d 'Resource name'
+complete -c mcpctl -n "__fish_seen_subcommand_from get describe delete edit approve; and not __mcpctl_needs_resource_type" -a '(__mcpctl_resource_names)' -d 'Resource name'
 
 # Helper: check if attach-server/detach-server already has a server argument
 function __mcpctl_needs_server_arg
@@ -196,22 +197,25 @@ complete -c mcpctl -n "__fish_seen_subcommand_from login" -l email -d 'Email add
 complete -c mcpctl -n "__fish_seen_subcommand_from login" -l password -d 'Password' -x
 
 # config subcommands
-set -l config_cmds view set path reset claude claude-generate impersonate
+set -l config_cmds view set path reset claude claude-generate setup impersonate
 complete -c mcpctl -n "__fish_seen_subcommand_from config; and not __fish_seen_subcommand_from $config_cmds" -a view -d 'Show configuration'
 complete -c mcpctl -n "__fish_seen_subcommand_from config; and not __fish_seen_subcommand_from $config_cmds" -a set -d 'Set a config value'
 complete -c mcpctl -n "__fish_seen_subcommand_from config; and not __fish_seen_subcommand_from $config_cmds" -a path -d 'Show config file path'
 complete -c mcpctl -n "__fish_seen_subcommand_from config; and not __fish_seen_subcommand_from $config_cmds" -a reset -d 'Reset to defaults'
 complete -c mcpctl -n "__fish_seen_subcommand_from config; and not __fish_seen_subcommand_from $config_cmds" -a claude -d 'Generate .mcp.json for project'
+complete -c mcpctl -n "__fish_seen_subcommand_from config; and not __fish_seen_subcommand_from $config_cmds" -a setup -d 'Configure LLM provider'
 complete -c mcpctl -n "__fish_seen_subcommand_from config; and not __fish_seen_subcommand_from $config_cmds" -a impersonate -d 'Impersonate a user'
 
 # create subcommands
-set -l create_cmds server secret project user group rbac
+set -l create_cmds server secret project user group rbac prompt promptrequest
 complete -c mcpctl -n "__fish_seen_subcommand_from create; and not __fish_seen_subcommand_from $create_cmds" -a server -d 'Create a server'
 complete -c mcpctl -n "__fish_seen_subcommand_from create; and not __fish_seen_subcommand_from $create_cmds" -a secret -d 'Create a secret'
 complete -c mcpctl -n "__fish_seen_subcommand_from create; and not __fish_seen_subcommand_from $create_cmds" -a project -d 'Create a project'
 complete -c mcpctl -n "__fish_seen_subcommand_from create; and not __fish_seen_subcommand_from $create_cmds" -a user -d 'Create a user'
 complete -c mcpctl -n "__fish_seen_subcommand_from create; and not __fish_seen_subcommand_from $create_cmds" -a group -d 'Create a group'
 complete -c mcpctl -n "__fish_seen_subcommand_from create; and not __fish_seen_subcommand_from $create_cmds" -a rbac -d 'Create an RBAC binding'
+complete -c mcpctl -n "__fish_seen_subcommand_from create; and not __fish_seen_subcommand_from $create_cmds" -a prompt -d 'Create an approved prompt'
+complete -c mcpctl -n "__fish_seen_subcommand_from create; and not __fish_seen_subcommand_from $create_cmds" -a promptrequest -d 'Create a prompt request'
 
 # logs options
 complete -c mcpctl -n "__fish_seen_subcommand_from logs" -l tail -d 'Number of lines' -x
@@ -227,6 +231,9 @@ complete -c mcpctl -n "__fish_seen_subcommand_from restore" -s i -l input -d 'In
 complete -c mcpctl -n "__fish_seen_subcommand_from restore" -s p -l password -d 'Decryption password' -x
 complete -c mcpctl -n "__fish_seen_subcommand_from restore" -s c -l conflict -d 'Conflict strategy' -xa 'skip overwrite fail'
 
+# approve: first arg is resource type (promptrequest only), second is name
+complete -c mcpctl -n "__fish_seen_subcommand_from approve; and __mcpctl_needs_resource_type" -a 'promptrequest' -d 'Resource type'
+
 # apply takes a file
 complete -c mcpctl -n "__fish_seen_subcommand_from apply" -s f -l file -d 'Configuration file' -rF
 complete -c mcpctl -n "__fish_seen_subcommand_from apply" -F

src/cli/src/commands/config-setup.ts

@@ -1,11 +1,15 @@
 import { Command } from 'commander';
 import http from 'node:http';
 import https from 'node:https';
+import { execFile } from 'node:child_process';
+import { promisify } from 'node:util';
 import { loadConfig, saveConfig } from '../config/index.js';
 import type { ConfigLoaderDeps, McpctlConfig, LlmConfig, LlmProviderName } from '../config/index.js';
 import type { SecretStore } from '@mcpctl/shared';
 import { createSecretStore } from '@mcpctl/shared';
 
+const execFileAsync = promisify(execFile);
+
 export interface ConfigSetupPrompt {
   select<T>(message: string, choices: Array<{ name: string; value: T; description?: string }>): Promise<T>;
   input(message: string, defaultValue?: string): Promise<string>;
@@ -19,6 +23,7 @@ export interface ConfigSetupDeps {
   log: (...args: string[]) => void;
   prompt: ConfigSetupPrompt;
   fetchModels: (url: string, path: string) => Promise<string[]>;
+  whichBinary: (name: string) => Promise<string | null>;
 }
 
 interface ProviderChoice {
@@ -130,6 +135,16 @@ const defaultPrompt: ConfigSetupPrompt = {
   confirm: defaultConfirm,
 };
 
+async function defaultWhichBinary(name: string): Promise<string | null> {
+  try {
+    const { stdout } = await execFileAsync('which', [name], { timeout: 3000 });
+    const path = stdout.trim();
+    return path || null;
+  } catch {
+    return null;
+  }
+}
+
 export function createConfigSetupCommand(deps?: Partial<ConfigSetupDeps>): Command {
   return new Command('setup')
     .description('Interactive LLM provider setup wizard')
@@ -138,6 +153,7 @@ export function createConfigSetupCommand(deps?: Partial<ConfigSetupDeps>): Comma
       const log = deps?.log ?? ((...args: string[]) => console.log(...args));
       const prompt = deps?.prompt ?? defaultPrompt;
       const fetchModels = deps?.fetchModels ?? defaultFetchModels;
+      const whichBinary = deps?.whichBinary ?? defaultWhichBinary;
       const secretStore = deps?.secretStore ?? await createSecretStore();
 
       const config = loadConfig(configDeps);
@@ -164,7 +180,7 @@ export function createConfigSetupCommand(deps?: Partial<ConfigSetupDeps>): Comma
 
       switch (provider) {
         case 'gemini-cli':
-          llmConfig = await setupGeminiCli(prompt, currentLlm);
+          llmConfig = await setupGeminiCli(prompt, log, whichBinary, currentLlm);
           break;
         case 'ollama':
           llmConfig = await setupOllama(prompt, fetchModels, currentLlm);
@@ -192,7 +208,12 @@ export function createConfigSetupCommand(deps?: Partial<ConfigSetupDeps>): Comma
     });
 }
 
-async function setupGeminiCli(prompt: ConfigSetupPrompt, current?: LlmConfig): Promise<LlmConfig> {
+async function setupGeminiCli(
+  prompt: ConfigSetupPrompt,
+  log: (...args: string[]) => void,
+  whichBinary: (name: string) => Promise<string | null>,
+  current?: LlmConfig,
+): Promise<LlmConfig> {
   const model = await prompt.select<string>('Select model:', [
     ...GEMINI_MODELS.map((m) => ({
       name: m === current?.model ? `${m} (current)` : m,
@@ -205,10 +226,17 @@ async function setupGeminiCli(prompt: ConfigSetupPrompt, current?: LlmConfig): P
     ? await prompt.input('Model name:', current?.model)
     : model;
 
-  const customBinary = await prompt.confirm('Use custom binary path?', false);
-  const binaryPath = customBinary
-    ? await prompt.input('Binary path:', current?.binaryPath ?? 'gemini')
-    : undefined;
+  // Auto-detect gemini binary path
+  let binaryPath: string | undefined;
+  const detected = await whichBinary('gemini');
+  if (detected) {
+    log(`Found gemini at: ${detected}`);
+    binaryPath = detected;
+  } else {
+    log('Warning: gemini binary not found in PATH');
+    const manualPath = await prompt.input('Binary path (or install with: npm i -g @google/gemini-cli):');
+    if (manualPath) binaryPath = manualPath;
+  }
 
   return { provider: 'gemini-cli', model: finalModel, binaryPath };
 }

src/cli/src/commands/create.ts

@@ -196,8 +196,6 @@ export function createCreateCommand(deps: CreateCommandDeps): Command {
     .argument('<name>', 'Project name')
     .option('-d, --description <text>', 'Project description', '')
     .option('--proxy-mode <mode>', 'Proxy mode (direct, filtered)')
-    .option('--proxy-mode-llm-provider <name>', 'LLM provider name (for filtered proxy mode)')
-    .option('--proxy-mode-llm-model <name>', 'LLM model name (for filtered proxy mode)')
     .option('--prompt <text>', 'Project-level prompt / instructions for the LLM')
     .option('--server <name>', 'Server name (repeat for multiple)', collect, [])
     .option('--force', 'Update if already exists')
@@ -208,8 +206,6 @@ export function createCreateCommand(deps: CreateCommandDeps): Command {
         proxyMode: opts.proxyMode ?? 'direct',
       };
       if (opts.prompt) body.prompt = opts.prompt;
-      if (opts.proxyModeLlmProvider) body.llmProvider = opts.proxyModeLlmProvider;
-      if (opts.proxyModeLlmModel) body.llmModel = opts.proxyModeLlmModel;
       if (opts.server.length > 0) body.servers = opts.server;
 
       try {
@@ -379,5 +375,31 @@ export function createCreateCommand(deps: CreateCommandDeps): Command {
       log(`prompt '${prompt.name}' created (id: ${prompt.id})`);
     });
 
+  // --- create promptrequest ---
+  cmd.command('promptrequest')
+    .description('Create a prompt request (pending proposal that needs approval)')
+    .argument('<name>', 'Prompt request name (lowercase alphanumeric with hyphens)')
+    .requiredOption('--project <name>', 'Project name (required)')
+    .option('--content <text>', 'Prompt content text')
+    .option('--content-file <path>', 'Read prompt content from file')
+    .action(async (name: string, opts) => {
+      let content = opts.content as string | undefined;
+      if (opts.contentFile) {
+        const fs = await import('node:fs/promises');
+        content = await fs.readFile(opts.contentFile as string, 'utf-8');
+      }
+      if (!content) {
+        throw new Error('--content or --content-file is required');
+      }
+
+      const projectName = opts.project as string;
+      const pr = await client.post<{ id: string; name: string }>(
+        `/api/v1/projects/${encodeURIComponent(projectName)}/promptrequests`,
+        { name, content },
+      );
+      log(`prompt request '${pr.name}' created (id: ${pr.id})`);
+      log(`  approve with: mcpctl approve promptrequest ${pr.name}`);
+    });
+
   return cmd;
 }

src/cli/src/commands/status.ts

@@ -7,11 +7,24 @@ import type { CredentialsDeps } from '../auth/index.js';
 import { formatJson, formatYaml } from '../formatters/index.js';
 import { APP_VERSION } from '@mcpctl/shared';
 
+// ANSI helpers
+const GREEN = '\x1b[32m';
+const RED = '\x1b[31m';
+const DIM = '\x1b[2m';
+const RESET = '\x1b[0m';
+const CLEAR_LINE = '\x1b[2K\r';
+
 export interface StatusCommandDeps {
   configDeps: Partial<ConfigLoaderDeps>;
   credentialsDeps: Partial<CredentialsDeps>;
   log: (...args: string[]) => void;
+  write: (text: string) => void;
   checkHealth: (url: string) => Promise<boolean>;
+  /** Check LLM health via mcplocal's /llm/health endpoint */
+  checkLlm: (mcplocalUrl: string) => Promise<string>;
+  /** Fetch available models from mcplocal's /llm/models endpoint */
+  fetchModels: (mcplocalUrl: string) => Promise<string[]>;
+  isTTY: boolean;
 }
 
 function defaultCheckHealth(url: string): Promise<boolean> {
@@ -28,15 +41,71 @@ function defaultCheckHealth(url: string): Promise<boolean> {
   });
 }
 
+/**
+ * Check LLM health by querying mcplocal's /llm/health endpoint.
+ * This tests the actual provider running inside the daemon (uses persistent ACP for gemini, etc.)
+ */
+function defaultCheckLlm(mcplocalUrl: string): Promise<string> {
+  return new Promise((resolve) => {
+    const req = http.get(`${mcplocalUrl}/llm/health`, { timeout: 30000 }, (res) => {
+      const chunks: Buffer[] = [];
+      res.on('data', (chunk: Buffer) => chunks.push(chunk));
+      res.on('end', () => {
+        try {
+          const body = JSON.parse(Buffer.concat(chunks).toString('utf-8')) as { status: string; error?: string };
+          if (body.status === 'ok') {
+            resolve('ok');
+          } else if (body.status === 'not configured') {
+            resolve('not configured');
+          } else if (body.error) {
+            resolve(body.error.slice(0, 80));
+          } else {
+            resolve(body.status);
+          }
+        } catch {
+          resolve('invalid response');
+        }
+      });
+    });
+    req.on('error', () => resolve('mcplocal unreachable'));
+    req.on('timeout', () => { req.destroy(); resolve('timeout'); });
+  });
+}
+
+function defaultFetchModels(mcplocalUrl: string): Promise<string[]> {
+  return new Promise((resolve) => {
+    const req = http.get(`${mcplocalUrl}/llm/models`, { timeout: 5000 }, (res) => {
+      const chunks: Buffer[] = [];
+      res.on('data', (chunk: Buffer) => chunks.push(chunk));
+      res.on('end', () => {
+        try {
+          const body = JSON.parse(Buffer.concat(chunks).toString('utf-8')) as { models?: string[] };
+          resolve(body.models ?? []);
+        } catch {
+          resolve([]);
+        }
+      });
+    });
+    req.on('error', () => resolve([]));
+    req.on('timeout', () => { req.destroy(); resolve([]); });
+  });
+}
+
+const SPINNER_FRAMES = ['⠋', '⠙', '⠹', '⠸', '⠼', '⠴', '⠦', '⠧', '⠇', '⠏'];
+
 const defaultDeps: StatusCommandDeps = {
   configDeps: {},
   credentialsDeps: {},
   log: (...args) => console.log(...args),
+  write: (text) => process.stdout.write(text),
   checkHealth: defaultCheckHealth,
+  checkLlm: defaultCheckLlm,
+  fetchModels: defaultFetchModels,
+  isTTY: process.stdout.isTTY ?? false,
 };
 
 export function createStatusCommand(deps?: Partial<StatusCommandDeps>): Command {
-  const { configDeps, credentialsDeps, log, checkHealth } = { ...defaultDeps, ...deps };
+  const { configDeps, credentialsDeps, log, write, checkHealth, checkLlm, fetchModels, isTTY } = { ...defaultDeps, ...deps };
 
   return new Command('status')
     .description('Show mcpctl status and connectivity')
@@ -45,13 +114,20 @@ export function createStatusCommand(deps?: Partial<StatusCommandDeps>): Command
       const config = loadConfig(configDeps);
       const creds = loadCredentials(credentialsDeps);
 
-      const [mcplocalReachable, mcpdReachable] = await Promise.all([
+      const llmLabel = config.llm && config.llm.provider !== 'none'
+        ? `${config.llm.provider}${config.llm.model ? ` / ${config.llm.model}` : ''}`
+        : null;
+
+      if (opts.output !== 'table') {
+        // JSON/YAML: run everything in parallel, wait, output at once
+        const [mcplocalReachable, mcpdReachable, llmStatus] = await Promise.all([
           checkHealth(config.mcplocalUrl),
           checkHealth(config.mcpdUrl),
+          llmLabel ? checkLlm(config.mcplocalUrl) : Promise.resolve(null),
         ]);
 
-      const llm = config.llm && config.llm.provider !== 'none'
-        ? `${config.llm.provider}${config.llm.model ? ` / ${config.llm.model}` : ''}`
+        const llm = llmLabel
+          ? llmStatus === 'ok' ? llmLabel : `${llmLabel} (${llmStatus})`
           : null;
 
         const status = {
@@ -64,21 +140,65 @@ export function createStatusCommand(deps?: Partial<StatusCommandDeps>): Command
           registries: config.registries,
           outputFormat: config.outputFormat,
           llm,
+          llmStatus,
         };
 
-      if (opts.output === 'json') {
-        log(formatJson(status));
-      } else if (opts.output === 'yaml') {
-        log(formatYaml(status));
-      } else {
-        log(`mcpctl v${status.version}`);
-        log(`mcplocal:   ${status.mcplocalUrl} (${mcplocalReachable ? 'connected' : 'unreachable'})`);
-        log(`mcpd:       ${status.mcpdUrl} (${mcpdReachable ? 'connected' : 'unreachable'})`);
-        log(`Auth:       ${creds ? `logged in as ${creds.user}` : 'not logged in'}`);
-        log(`Registries: ${status.registries.join(', ')}`);
-        log(`Output:     ${status.outputFormat}`);
-        log(`LLM:        ${status.llm ?? "not configured (run 'mcpctl config setup')"}`);
+        log(opts.output === 'json' ? formatJson(status) : formatYaml(status));
+        return;
+      }
 
+      // Table format: print lines progressively, LLM last with spinner
+
+      // Fast health checks first
+      const [mcplocalReachable, mcpdReachable] = await Promise.all([
+        checkHealth(config.mcplocalUrl),
+        checkHealth(config.mcpdUrl),
+      ]);
+
+      log(`mcpctl v${APP_VERSION}`);
+      log(`mcplocal:   ${config.mcplocalUrl} (${mcplocalReachable ? 'connected' : 'unreachable'})`);
+      log(`mcpd:       ${config.mcpdUrl} (${mcpdReachable ? 'connected' : 'unreachable'})`);
+      log(`Auth:       ${creds ? `logged in as ${creds.user}` : 'not logged in'}`);
+      log(`Registries: ${config.registries.join(', ')}`);
+      log(`Output:     ${config.outputFormat}`);
+
+      if (!llmLabel) {
+        log(`LLM:        not configured (run 'mcpctl config setup')`);
+        return;
+      }
+
+      // LLM check with spinner — queries mcplocal's /llm/health endpoint
+      const llmPromise = checkLlm(config.mcplocalUrl);
+
+      if (isTTY) {
+        let frame = 0;
+        const interval = setInterval(() => {
+          write(`${CLEAR_LINE}LLM:        ${llmLabel} ${DIM}${SPINNER_FRAMES[frame % SPINNER_FRAMES.length]} checking...${RESET}`);
+          frame++;
+        }, 80);
+
+        const llmStatus = await llmPromise;
+        clearInterval(interval);
+
+        if (llmStatus === 'ok' || llmStatus === 'ok (key stored)') {
+          write(`${CLEAR_LINE}LLM:        ${llmLabel} ${GREEN}✓ ${llmStatus}${RESET}\n`);
+        } else {
+          write(`${CLEAR_LINE}LLM:        ${llmLabel} ${RED}✗ ${llmStatus}${RESET}\n`);
+        }
+      } else {
+        // Non-TTY: no spinner, just wait and print
+        const llmStatus = await llmPromise;
+        if (llmStatus === 'ok' || llmStatus === 'ok (key stored)') {
+          log(`LLM:        ${llmLabel} ✓ ${llmStatus}`);
+        } else {
+          log(`LLM:        ${llmLabel} ✗ ${llmStatus}`);
+        }
+      }
+
+      // Show available models (non-blocking, best effort)
+      const models = await fetchModels(config.mcplocalUrl);
+      if (models.length > 0) {
+        log(`${DIM}            Available: ${models.join(', ')}${RESET}`);
       }
     });
 }

src/cli/tests/commands/config-setup.test.ts

@@ -42,6 +42,7 @@ function buildDeps(overrides: {
   secrets?: Record<string, string>;
   answers?: unknown[];
   fetchModels?: ConfigSetupDeps['fetchModels'];
+  whichBinary?: ConfigSetupDeps['whichBinary'];
 } = {}): ConfigSetupDeps {
   return {
     configDeps: { configDir: tempDir },
@@ -49,6 +50,7 @@ function buildDeps(overrides: {
     log: (...args: string[]) => logs.push(args.join(' ')),
     prompt: mockPrompt(overrides.answers ?? []),
     fetchModels: overrides.fetchModels ?? vi.fn(async () => []),
+    whichBinary: overrides.whichBinary ?? vi.fn(async () => '/usr/bin/gemini'),
   };
 }
 
@@ -76,26 +78,49 @@ describe('config setup wizard', () => {
   });
 
   describe('provider: gemini-cli', () => {
-    it('saves gemini-cli with selected model', async () => {
-      // Answers: select provider, select model, confirm custom binary=false
-      const deps = buildDeps({ answers: ['gemini-cli', 'gemini-2.5-flash', false] });
+    it('auto-detects binary path and saves config', async () => {
+      // Answers: select provider, select model (no binary prompt — auto-detected)
+      const deps = buildDeps({
+        answers: ['gemini-cli', 'gemini-2.5-flash'],
+        whichBinary: vi.fn(async () => '/home/user/.npm-global/bin/gemini'),
+      });
       await runSetup(deps);
 
       const config = readConfig();
-      expect((config.llm as Record<string, unknown>).provider).toBe('gemini-cli');
-      expect((config.llm as Record<string, unknown>).model).toBe('gemini-2.5-flash');
+      const llm = config.llm as Record<string, unknown>;
+      expect(llm.provider).toBe('gemini-cli');
+      expect(llm.model).toBe('gemini-2.5-flash');
+      expect(llm.binaryPath).toBe('/home/user/.npm-global/bin/gemini');
+      expect(logs.some((l) => l.includes('Found gemini at'))).toBe(true);
       cleanup();
     });
 
-    it('saves gemini-cli with custom model and binary path', async () => {
-      // Answers: select provider, select custom, enter model name, confirm custom binary=true, enter path
-      const deps = buildDeps({ answers: ['gemini-cli', '__custom__', 'gemini-3.0-flash', true, '/opt/gemini'] });
+    it('prompts for manual path when binary not found', async () => {
+      // Answers: select provider, select model, enter manual path
+      const deps = buildDeps({
+        answers: ['gemini-cli', 'gemini-2.5-flash', '/opt/gemini'],
+        whichBinary: vi.fn(async () => null),
+      });
+      await runSetup(deps);
+
+      const config = readConfig();
+      const llm = config.llm as Record<string, unknown>;
+      expect(llm.binaryPath).toBe('/opt/gemini');
+      expect(logs.some((l) => l.includes('not found'))).toBe(true);
+      cleanup();
+    });
+
+    it('saves gemini-cli with custom model', async () => {
+      // Answers: select provider, select custom, enter model name
+      const deps = buildDeps({
+        answers: ['gemini-cli', '__custom__', 'gemini-3.0-flash'],
+        whichBinary: vi.fn(async () => '/usr/bin/gemini'),
+      });
       await runSetup(deps);
 
       const config = readConfig();
       const llm = config.llm as Record<string, unknown>;
       expect(llm.model).toBe('gemini-3.0-flash');
-      expect(llm.binaryPath).toBe('/opt/gemini');
       cleanup();
     });
   });
@@ -250,7 +275,7 @@ describe('config setup wizard', () => {
 
   describe('output messages', () => {
     it('shows restart instruction', async () => {
-      const deps = buildDeps({ answers: ['gemini-cli', 'gemini-2.5-flash', false] });
+      const deps = buildDeps({ answers: ['gemini-cli', 'gemini-2.5-flash'] });
       await runSetup(deps);
 
       expect(logs.some((l) => l.includes('systemctl --user restart mcplocal'))).toBe(true);
@@ -258,7 +283,7 @@ describe('config setup wizard', () => {
     });
 
     it('shows configured provider and model', async () => {
-      const deps = buildDeps({ answers: ['gemini-cli', 'gemini-2.5-flash', false] });
+      const deps = buildDeps({ answers: ['gemini-cli', 'gemini-2.5-flash'] });
       await runSetup(deps);
 
       expect(logs.some((l) => l.includes('gemini-cli') && l.includes('gemini-2.5-flash'))).toBe(true);

src/cli/tests/commands/project.test.ts

@@ -30,8 +30,6 @@ describe('project with new fields', () => {
         'project', 'smart-home',
         '-d', 'Smart home project',
         '--proxy-mode', 'filtered',
-        '--proxy-mode-llm-provider', 'gemini-cli',
-        '--proxy-mode-llm-model', 'gemini-2.0-flash',
         '--server', 'my-grafana',
         '--server', 'my-ha',
       ], { from: 'user' });
@@ -40,8 +38,6 @@ describe('project with new fields', () => {
         name: 'smart-home',
         description: 'Smart home project',
         proxyMode: 'filtered',
-        llmProvider: 'gemini-cli',
-        llmModel: 'gemini-2.0-flash',
         servers: ['my-grafana', 'my-ha'],
       }));
     });

src/cli/tests/commands/status.test.ts

@@ -3,19 +3,38 @@ import { mkdtempSync, rmSync } from 'node:fs';
 import { join } from 'node:path';
 import { tmpdir } from 'node:os';
 import { createStatusCommand } from '../../src/commands/status.js';
+import type { StatusCommandDeps } from '../../src/commands/status.js';
 import { saveConfig, DEFAULT_CONFIG } from '../../src/config/index.js';
 import { saveCredentials } from '../../src/auth/index.js';
 
 let tempDir: string;
 let output: string[];
+let written: string[];
 
 function log(...args: string[]) {
   output.push(args.join(' '));
 }
 
+function write(text: string) {
+  written.push(text);
+}
+
+function baseDeps(overrides?: Partial<StatusCommandDeps>): Partial<StatusCommandDeps> {
+  return {
+    configDeps: { configDir: tempDir },
+    credentialsDeps: { configDir: tempDir },
+    log,
+    write,
+    checkHealth: async () => true,
+    isTTY: false,
+    ...overrides,
+  };
+}
+
 beforeEach(() => {
   tempDir = mkdtempSync(join(tmpdir(), 'mcpctl-status-test-'));
   output = [];
+  written = [];
 });
 
 afterEach(() => {
@@ -24,12 +43,7 @@ afterEach(() => {
 
 describe('status command', () => {
   it('shows status in table format', async () => {
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps());
     await cmd.parseAsync([], { from: 'user' });
     const out = output.join('\n');
     expect(out).toContain('mcpctl v');
@@ -39,46 +53,26 @@ describe('status command', () => {
   });
 
   it('shows unreachable when daemons are down', async () => {
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => false,
-    });
+    const cmd = createStatusCommand(baseDeps({ checkHealth: async () => false }));
     await cmd.parseAsync([], { from: 'user' });
     expect(output.join('\n')).toContain('unreachable');
   });
 
   it('shows not logged in when no credentials', async () => {
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps());
     await cmd.parseAsync([], { from: 'user' });
     expect(output.join('\n')).toContain('not logged in');
   });
 
   it('shows logged in user when credentials exist', async () => {
     saveCredentials({ token: 'tok', mcpdUrl: 'http://x:3100', user: 'alice@example.com' }, { configDir: tempDir });
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps());
     await cmd.parseAsync([], { from: 'user' });
     expect(output.join('\n')).toContain('logged in as alice@example.com');
   });
 
   it('shows status in JSON format', async () => {
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps());
     await cmd.parseAsync(['-o', 'json'], { from: 'user' });
     const parsed = JSON.parse(output[0]) as Record<string, unknown>;
     expect(parsed['version']).toBe('0.1.0');
@@ -87,12 +81,7 @@ describe('status command', () => {
   });
 
   it('shows status in YAML format', async () => {
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => false,
-    });
+    const cmd = createStatusCommand(baseDeps({ checkHealth: async () => false }));
     await cmd.parseAsync(['-o', 'yaml'], { from: 'user' });
     expect(output[0]).toContain('mcplocalReachable: false');
   });
@@ -100,15 +89,12 @@ describe('status command', () => {
   it('checks correct URLs from config', async () => {
     saveConfig({ ...DEFAULT_CONFIG, mcplocalUrl: 'http://local:3200', mcpdUrl: 'http://remote:3100' }, { configDir: tempDir });
     const checkedUrls: string[] = [];
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
+    const cmd = createStatusCommand(baseDeps({
       checkHealth: async (url) => {
         checkedUrls.push(url);
         return false;
       },
-    });
+    }));
     await cmd.parseAsync([], { from: 'user' });
     expect(checkedUrls).toContain('http://local:3200');
     expect(checkedUrls).toContain('http://remote:3100');
@@ -116,24 +102,14 @@ describe('status command', () => {
 
   it('shows registries from config', async () => {
     saveConfig({ ...DEFAULT_CONFIG, registries: ['official'] }, { configDir: tempDir });
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps());
     await cmd.parseAsync([], { from: 'user' });
     expect(output.join('\n')).toContain('official');
     expect(output.join('\n')).not.toContain('glama');
   });
 
   it('shows LLM not configured hint when no LLM is set', async () => {
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps());
     await cmd.parseAsync([], { from: 'user' });
     const out = output.join('\n');
     expect(out).toContain('LLM:');
@@ -141,54 +117,85 @@ describe('status command', () => {
     expect(out).toContain('mcpctl config setup');
   });
 
-  it('shows configured LLM provider and model', async () => {
+  it('shows green check when LLM is healthy (non-TTY)', async () => {
     saveConfig({ ...DEFAULT_CONFIG, llm: { provider: 'anthropic', model: 'claude-haiku-3-5-20241022' } }, { configDir: tempDir });
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps({ checkLlm: async () => 'ok' }));
     await cmd.parseAsync([], { from: 'user' });
     const out = output.join('\n');
-    expect(out).toContain('LLM:');
     expect(out).toContain('anthropic / claude-haiku-3-5-20241022');
+    expect(out).toContain('✓ ok');
+  });
+
+  it('shows red cross when LLM check fails (non-TTY)', async () => {
+    saveConfig({ ...DEFAULT_CONFIG, llm: { provider: 'gemini-cli', model: 'gemini-2.5-flash' } }, { configDir: tempDir });
+    const cmd = createStatusCommand(baseDeps({ checkLlm: async () => 'not authenticated' }));
+    await cmd.parseAsync([], { from: 'user' });
+    const out = output.join('\n');
+    expect(out).toContain('✗ not authenticated');
+  });
+
+  it('shows error message from mcplocal', async () => {
+    saveConfig({ ...DEFAULT_CONFIG, llm: { provider: 'gemini-cli', model: 'gemini-2.5-flash' } }, { configDir: tempDir });
+    const cmd = createStatusCommand(baseDeps({ checkLlm: async () => 'binary not found' }));
+    await cmd.parseAsync([], { from: 'user' });
+    expect(output.join('\n')).toContain('✗ binary not found');
+  });
+
+  it('queries mcplocal URL for LLM health', async () => {
+    saveConfig({ ...DEFAULT_CONFIG, mcplocalUrl: 'http://custom:9999', llm: { provider: 'gemini-cli', model: 'gemini-2.5-flash' } }, { configDir: tempDir });
+    let queriedUrl = '';
+    const cmd = createStatusCommand(baseDeps({
+      checkLlm: async (url) => { queriedUrl = url; return 'ok'; },
+    }));
+    await cmd.parseAsync([], { from: 'user' });
+    expect(queriedUrl).toBe('http://custom:9999');
+  });
+
+  it('uses spinner on TTY and writes final result', async () => {
+    saveConfig({ ...DEFAULT_CONFIG, llm: { provider: 'gemini-cli', model: 'gemini-2.5-flash' } }, { configDir: tempDir });
+    const cmd = createStatusCommand(baseDeps({
+      isTTY: true,
+      checkLlm: async () => 'ok',
+    }));
+    await cmd.parseAsync([], { from: 'user' });
+    // On TTY, the final LLM line goes through write(), not log()
+    const finalWrite = written[written.length - 1];
+    expect(finalWrite).toContain('gemini-cli / gemini-2.5-flash');
+    expect(finalWrite).toContain('✓ ok');
+  });
+
+  it('uses spinner on TTY and shows failure', async () => {
+    saveConfig({ ...DEFAULT_CONFIG, llm: { provider: 'gemini-cli', model: 'gemini-2.5-flash' } }, { configDir: tempDir });
+    const cmd = createStatusCommand(baseDeps({
+      isTTY: true,
+      checkLlm: async () => 'not authenticated',
+    }));
+    await cmd.parseAsync([], { from: 'user' });
+    const finalWrite = written[written.length - 1];
+    expect(finalWrite).toContain('✗ not authenticated');
   });
 
   it('shows not configured when LLM provider is none', async () => {
     saveConfig({ ...DEFAULT_CONFIG, llm: { provider: 'none' } }, { configDir: tempDir });
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps());
     await cmd.parseAsync([], { from: 'user' });
     expect(output.join('\n')).toContain('not configured');
   });
 
-  it('includes llm field in JSON output', async () => {
+  it('includes llm and llmStatus in JSON output', async () => {
     saveConfig({ ...DEFAULT_CONFIG, llm: { provider: 'gemini-cli', model: 'gemini-2.5-flash' } }, { configDir: tempDir });
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps({ checkLlm: async () => 'ok' }));
     await cmd.parseAsync(['-o', 'json'], { from: 'user' });
     const parsed = JSON.parse(output[0]) as Record<string, unknown>;
     expect(parsed['llm']).toBe('gemini-cli / gemini-2.5-flash');
+    expect(parsed['llmStatus']).toBe('ok');
   });
 
   it('includes null llm in JSON output when not configured', async () => {
-    const cmd = createStatusCommand({
-      configDeps: { configDir: tempDir },
-      credentialsDeps: { configDir: tempDir },
-      log,
-      checkHealth: async () => true,
-    });
+    const cmd = createStatusCommand(baseDeps());
     await cmd.parseAsync(['-o', 'json'], { from: 'user' });
     const parsed = JSON.parse(output[0]) as Record<string, unknown>;
     expect(parsed['llm']).toBeNull();
+    expect(parsed['llmStatus']).toBeNull();
   });
 });

src/cli/tests/completions.test.ts

@@ -15,7 +15,7 @@ describe('fish completions', () => {
   });
 
   it('does not offer resource types without __mcpctl_needs_resource_type guard', () => {
-    const resourceTypes = ['servers', 'instances', 'secrets', 'templates', 'projects', 'users', 'groups', 'rbac'];
+    const resourceTypes = ['servers', 'instances', 'secrets', 'templates', 'projects', 'users', 'groups', 'rbac', 'prompts', 'promptrequests'];
     const lines = fishFile.split('\n').filter((l) => l.startsWith('complete '));
 
     for (const line of lines) {

src/mcplocal/src/discovery.ts

@@ -48,6 +48,33 @@ export async function refreshProjectUpstreams(
   return syncUpstreams(router, mcpdClient, servers);
 }
 
+/**
+ * Fetch a project's LLM config (llmProvider, llmModel) from mcpd.
+ * These are the project-level "recommendations" — local overrides take priority.
+ */
+export interface ProjectLlmConfig {
+  llmProvider?: string;
+  llmModel?: string;
+}
+
+export async function fetchProjectLlmConfig(
+  mcpdClient: McpdClient,
+  projectName: string,
+): Promise<ProjectLlmConfig> {
+  try {
+    const project = await mcpdClient.get<{
+      llmProvider?: string;
+      llmModel?: string;
+    }>(`/api/v1/projects/${encodeURIComponent(projectName)}`);
+    const config: ProjectLlmConfig = {};
+    if (project.llmProvider) config.llmProvider = project.llmProvider;
+    if (project.llmModel) config.llmModel = project.llmModel;
+    return config;
+  } catch {
+    return {};
+  }
+}
+
 /** Shared sync logic: reconcile a router's upstreams with a server list. */
 function syncUpstreams(router: McpRouter, mcpdClient: McpdClient, servers: McpdServer[]): string[] {
   const registered: string[] = [];

src/mcplocal/src/http/config.ts

@@ -44,21 +44,54 @@ export interface LlmFileConfig {
   binaryPath?: string;
 }
 
+export interface ProjectLlmOverride {
+  model?: string;
+  provider?: string;
+}
+
+interface McpctlConfig {
+  llm?: LlmFileConfig;
+  projects?: Record<string, { llm?: ProjectLlmOverride }>;
+}
+
+/** Cached config for the process lifetime (reloaded on SIGHUP if needed). */
+let cachedConfig: McpctlConfig | null = null;
+
+function loadFullConfig(): McpctlConfig {
+  if (cachedConfig) return cachedConfig;
+  try {
+    const configPath = join(homedir(), '.mcpctl', 'config.json');
+    if (!existsSync(configPath)) return {};
+    const raw = readFileSync(configPath, 'utf-8');
+    cachedConfig = JSON.parse(raw) as McpctlConfig;
+    return cachedConfig;
+  } catch {
+    return {};
+  }
+}
+
 /**
  * Load LLM configuration from ~/.mcpctl/config.json.
  * Returns undefined if no LLM section is configured.
  */
 export function loadLlmConfig(): LlmFileConfig | undefined {
-  try {
-    const configPath = join(homedir(), '.mcpctl', 'config.json');
-    if (!existsSync(configPath)) return undefined;
-    const raw = readFileSync(configPath, 'utf-8');
-    const parsed = JSON.parse(raw) as { llm?: LlmFileConfig };
-    if (!parsed.llm?.provider || parsed.llm.provider === 'none') return undefined;
-    return parsed.llm;
-  } catch {
-    return undefined;
-  }
+  const config = loadFullConfig();
+  if (!config.llm?.provider || config.llm.provider === 'none') return undefined;
+  return config.llm;
+}
+
+/**
+ * Load per-project LLM override from ~/.mcpctl/config.json.
+ * Returns the project-specific model/provider override, or undefined.
+ */
+export function loadProjectLlmOverride(projectName: string): ProjectLlmOverride | undefined {
+  const config = loadFullConfig();
+  return config.projects?.[projectName]?.llm;
+}
+
+/** Reset cached config (for testing). */
+export function resetConfigCache(): void {
+  cachedConfig = null;
 }
 
 export function loadHttpConfig(env: Record<string, string | undefined> = process.env): HttpConfig {

src/mcplocal/src/http/project-mcp-endpoint.ts

@@ -13,7 +13,8 @@ import { StreamableHTTPServerTransport } from '@modelcontextprotocol/sdk/server/
 import type { JSONRPCMessage } from '@modelcontextprotocol/sdk/types.js';
 import { McpRouter } from '../router.js';
 import { ResponsePaginator } from '../llm/pagination.js';
-import { refreshProjectUpstreams } from '../discovery.js';
+import { refreshProjectUpstreams, fetchProjectLlmConfig } from '../discovery.js';
+import { loadProjectLlmOverride } from './config.js';
 import type { McpdClient } from './mcpd-client.js';
 import type { ProviderRegistry } from '../providers/registry.js';
 import type { JsonRpcRequest } from '../types.js';
@@ -46,8 +47,13 @@ export function registerProjectMcpEndpoint(app: FastifyInstance, mcpdClient: Mcp
     const router = existing?.router ?? new McpRouter();
     await refreshProjectUpstreams(router, mcpdClient, projectName, authToken);
 
-    // Wire pagination support with LLM provider if configured
-    router.setPaginator(new ResponsePaginator(providerRegistry?.getActive() ?? null));
+    // Resolve project LLM model: local override → mcpd recommendation → global default
+    const localOverride = loadProjectLlmOverride(projectName);
+    const mcpdConfig = await fetchProjectLlmConfig(mcpdClient, projectName);
+    const resolvedModel = localOverride?.model ?? mcpdConfig.llmModel ?? undefined;
+
+    // Wire pagination support with LLM provider and project model override
+    router.setPaginator(new ResponsePaginator(providerRegistry ?? null, {}, resolvedModel));
 
     // Configure prompt resources with SA-scoped client for RBAC
     const saClient = mcpdClient.withHeaders({ 'X-Service-Account': `project:${projectName}` });

src/mcplocal/src/http/server.ts

@@ -81,6 +81,49 @@ export async function createHttpServer(
     reply.code(200).send({ status: 'ok' });
   });
 
+  // LLM health check — tests the active provider with a tiny prompt
+  app.get('/llm/health', async (_request, reply) => {
+    const provider = deps.providerRegistry?.getActive() ?? null;
+    if (!provider) {
+      reply.code(200).send({ status: 'not configured' });
+      return;
+    }
+    try {
+      const result = await provider.complete({
+        messages: [{ role: 'user', content: 'Respond with exactly: ok' }],
+        maxTokens: 10,
+      });
+      const ok = result.content.trim().toLowerCase().includes('ok');
+      reply.code(200).send({
+        status: ok ? 'ok' : 'unexpected response',
+        provider: provider.name,
+        response: result.content.trim().slice(0, 100),
+      });
+    } catch (err) {
+      const msg = (err as Error).message ?? String(err);
+      reply.code(200).send({
+        status: 'error',
+        provider: provider.name,
+        error: msg.slice(0, 200),
+      });
+    }
+  });
+
+  // LLM models — list available models from the active provider
+  app.get('/llm/models', async (_request, reply) => {
+    const provider = deps.providerRegistry?.getActive() ?? null;
+    if (!provider) {
+      reply.code(200).send({ models: [], provider: null });
+      return;
+    }
+    try {
+      const models = await provider.listModels();
+      reply.code(200).send({ models, provider: provider.name });
+    } catch {
+      reply.code(200).send({ models: [], provider: provider.name });
+    }
+  });
+
   // Proxy management routes to mcpd
   const mcpdClient = new McpdClient(config.mcpdUrl, config.mcpdToken);
   registerProxyRoutes(app, mcpdClient);

src/mcplocal/src/llm-config.ts

@@ -1,11 +1,16 @@
 import type { SecretStore } from '@mcpctl/shared';
 import type { LlmFileConfig } from './http/config.js';
 import { ProviderRegistry } from './providers/registry.js';
-import { GeminiCliProvider } from './providers/gemini-cli.js';
+import { GeminiAcpProvider } from './providers/gemini-acp.js';
 import { OllamaProvider } from './providers/ollama.js';
 import { AnthropicProvider } from './providers/anthropic.js';
 import { OpenAiProvider } from './providers/openai.js';
 import { DeepSeekProvider } from './providers/deepseek.js';
+import type { GeminiAcpConfig } from './providers/gemini-acp.js';
+import type { OllamaConfig } from './providers/ollama.js';
+import type { AnthropicConfig } from './providers/anthropic.js';
+import type { OpenAiConfig } from './providers/openai.js';
+import type { DeepSeekConfig } from './providers/deepseek.js';
 
 /**
  * Create a ProviderRegistry from user config + secret store.
@@ -19,19 +24,21 @@ export async function createProviderFromConfig(
   if (!config?.provider || config.provider === 'none') return registry;
 
   switch (config.provider) {
-    case 'gemini-cli':
-      registry.register(new GeminiCliProvider({
-        binaryPath: config.binaryPath,
-        defaultModel: config.model,
-      }));
+    case 'gemini-cli': {
+      const cfg: GeminiAcpConfig = {};
+      if (config.binaryPath) cfg.binaryPath = config.binaryPath;
+      if (config.model) cfg.defaultModel = config.model;
+      registry.register(new GeminiAcpProvider(cfg));
       break;
+    }
 
-    case 'ollama':
-      registry.register(new OllamaProvider({
-        baseUrl: config.url,
-        defaultModel: config.model,
-      }));
+    case 'ollama': {
+      const cfg: OllamaConfig = {};
+      if (config.url) cfg.baseUrl = config.url;
+      if (config.model) cfg.defaultModel = config.model;
+      registry.register(new OllamaProvider(cfg));
       break;
+    }
 
     case 'anthropic': {
       const apiKey = await secretStore.get('anthropic-api-key');
@@ -39,10 +46,9 @@ export async function createProviderFromConfig(
         process.stderr.write('Warning: Anthropic API key not found in secret store. Run "mcpctl config setup" to configure.\n');
         return registry;
       }
-      registry.register(new AnthropicProvider({
-        apiKey,
-        defaultModel: config.model,
-      }));
+      const cfg: AnthropicConfig = { apiKey };
+      if (config.model) cfg.defaultModel = config.model;
+      registry.register(new AnthropicProvider(cfg));
       break;
     }
 
@@ -52,11 +58,10 @@ export async function createProviderFromConfig(
         process.stderr.write('Warning: OpenAI API key not found in secret store. Run "mcpctl config setup" to configure.\n');
         return registry;
       }
-      registry.register(new OpenAiProvider({
-        apiKey,
-        baseUrl: config.url,
-        defaultModel: config.model,
-      }));
+      const cfg: OpenAiConfig = { apiKey };
+      if (config.url) cfg.baseUrl = config.url;
+      if (config.model) cfg.defaultModel = config.model;
+      registry.register(new OpenAiProvider(cfg));
       break;
     }
 
@@ -66,11 +71,10 @@ export async function createProviderFromConfig(
         process.stderr.write('Warning: DeepSeek API key not found in secret store. Run "mcpctl config setup" to configure.\n');
         return registry;
       }
-      registry.register(new DeepSeekProvider({
-        apiKey,
-        baseUrl: config.url,
-        defaultModel: config.model,
-      }));
+      const cfg: DeepSeekConfig = { apiKey };
+      if (config.url) cfg.baseUrl = config.url;
+      if (config.model) cfg.defaultModel = config.model;
+      registry.register(new DeepSeekProvider(cfg));
       break;
     }
 

src/mcplocal/src/llm/pagination.ts

@@ -105,6 +105,7 @@ export class ResponsePaginator {
   constructor(
     private providers: ProviderRegistry | null,
     config: Partial<PaginationConfig> = {},
+    private modelOverride?: string,
   ) {
     this.config = { ...DEFAULT_PAGINATION_CONFIG, ...config };
   }
@@ -129,7 +130,8 @@ export class ResponsePaginator {
 
     try {
       index = await this.generateSmartIndex(resultId, toolName, raw, pages);
-    } catch {
+    } catch (err) {
+      console.error(`[pagination] Smart index failed for ${toolName}, falling back to simple:`, err instanceof Error ? err.message : String(err));
       index = this.generateSimpleIndex(resultId, toolName, raw, pages);
     }
 
@@ -259,9 +261,12 @@ export class ResponsePaginator {
       ],
       maxTokens: this.config.indexMaxTokens,
       temperature: 0,
+      ...(this.modelOverride ? { model: this.modelOverride } : {}),
     });
 
-    const summaries = JSON.parse(result.content) as Array<{ page: number; summary: string }>;
+    // LLMs often wrap JSON in ```json ... ``` fences — strip them
+    const cleaned = result.content.replace(/^```(?:json)?\s*\n?/i, '').replace(/\n?```\s*$/i, '').trim();
+    const summaries = JSON.parse(cleaned) as Array<{ page: number; summary: string }>;
 
     return {
       resultId,

src/mcplocal/src/main.ts

@@ -139,6 +139,7 @@ export async function main(argv: string[] = process.argv): Promise<MainResult> {
     if (shuttingDown) return;
     shuttingDown = true;
 
+    providerRegistry.disposeAll();
     server.stop();
     if (httpServer) {
       await httpServer.close();

src/mcplocal/src/providers/acp-client.ts

@@ -0,0 +1,291 @@
+import { spawn, type ChildProcess } from 'node:child_process';
+import { createInterface, type Interface as ReadlineInterface } from 'node:readline';
+
+export interface AcpClientConfig {
+  binaryPath: string;
+  model: string;
+  /** Timeout for individual RPC requests in ms (default: 60000) */
+  requestTimeoutMs: number;
+  /** Timeout for process initialization in ms (default: 30000) */
+  initTimeoutMs: number;
+  /** Override spawn for testing */
+  spawn?: typeof spawn;
+}
+
+interface PendingRequest {
+  resolve: (result: unknown) => void;
+  reject: (err: Error) => void;
+  timer: ReturnType<typeof setTimeout>;
+}
+
+/**
+ * Low-level ACP (Agent Client Protocol) client.
+ * Manages a persistent `gemini --experimental-acp` subprocess and communicates
+ * via JSON-RPC 2.0 over NDJSON stdio.
+ *
+ * Pattern follows StdioUpstream: readline for parsing, pending request map with timeouts.
+ */
+export class AcpClient {
+  private process: ChildProcess | null = null;
+  private readline: ReadlineInterface | null = null;
+  private pendingRequests = new Map<number, PendingRequest>();
+  private nextId = 1;
+  private sessionId: string | null = null;
+  private ready = false;
+  private initPromise: Promise<void> | null = null;
+  private readonly config: AcpClientConfig;
+
+  /** Accumulates text chunks from session/update agent_message_chunk during a prompt. */
+  private activePromptChunks: string[] = [];
+
+  constructor(config: AcpClientConfig) {
+    this.config = config;
+  }
+
+  /** Ensure the subprocess is spawned and initialized. Idempotent and lazy. */
+  async ensureReady(): Promise<void> {
+    if (this.ready && this.process && !this.process.killed) return;
+
+    // If already initializing, wait for it
+    if (this.initPromise) return this.initPromise;
+
+    this.initPromise = this.doInit();
+    try {
+      await this.initPromise;
+    } catch (err) {
+      this.initPromise = null;
+      throw err;
+    }
+  }
+
+  /** Send a prompt and collect the streamed text response. */
+  async prompt(text: string): Promise<string> {
+    await this.ensureReady();
+
+    // Set up chunk accumulator
+    this.activePromptChunks = [];
+
+    const result = await this.sendRequest('session/prompt', {
+      sessionId: this.sessionId,
+      prompt: [{ type: 'text', text }],
+    }, this.config.requestTimeoutMs) as { stopReason: string };
+
+    const collected = this.activePromptChunks.join('');
+    this.activePromptChunks = [];
+
+    if (result.stopReason === 'refusal') {
+      throw new Error('Gemini refused to process the prompt');
+    }
+
+    return collected;
+  }
+
+  /** Kill the subprocess and clean up. */
+  dispose(): void {
+    this.cleanup();
+  }
+
+  /** Check if the subprocess is alive and initialized. */
+  get isAlive(): boolean {
+    return this.ready && this.process !== null && !this.process.killed;
+  }
+
+  // --- Private ---
+
+  private async doInit(): Promise<void> {
+    // Clean up any previous state
+    this.cleanup();
+
+    this.spawnProcess();
+    this.setupReadline();
+
+    // ACP handshake: initialize
+    await this.sendRequest('initialize', {
+      protocolVersion: 1,
+      clientCapabilities: {},
+      clientInfo: { name: 'mcpctl', version: '1.0.0' },
+    }, this.config.initTimeoutMs);
+
+    // ACP handshake: session/new
+    const sessionResult = await this.sendRequest('session/new', {
+      cwd: '/tmp',
+      mcpServers: [],
+    }, this.config.initTimeoutMs) as { sessionId: string };
+
+    this.sessionId = sessionResult.sessionId;
+    this.ready = true;
+  }
+
+  private spawnProcess(): void {
+    const spawnFn = this.config.spawn ?? spawn;
+    this.process = spawnFn(this.config.binaryPath, ['--experimental-acp'], {
+      stdio: ['pipe', 'pipe', 'pipe'],
+      env: process.env,
+    });
+
+    this.process.on('exit', () => {
+      this.ready = false;
+      this.initPromise = null;
+      this.sessionId = null;
+
+      // Reject all pending requests
+      for (const [id, pending] of this.pendingRequests) {
+        clearTimeout(pending.timer);
+        pending.reject(new Error('Gemini ACP process exited'));
+        this.pendingRequests.delete(id);
+      }
+    });
+
+    this.process.on('error', (err) => {
+      this.ready = false;
+      this.initPromise = null;
+
+      for (const [id, pending] of this.pendingRequests) {
+        clearTimeout(pending.timer);
+        pending.reject(err);
+        this.pendingRequests.delete(id);
+      }
+    });
+  }
+
+  private setupReadline(): void {
+    if (!this.process?.stdout) return;
+
+    this.readline = createInterface({ input: this.process.stdout });
+    this.readline.on('line', (line) => this.handleLine(line));
+  }
+
+  private handleLine(line: string): void {
+    let msg: Record<string, unknown>;
+    try {
+      msg = JSON.parse(line) as Record<string, unknown>;
+    } catch {
+      // Skip non-JSON lines (e.g., debug output on stdout)
+      return;
+    }
+
+    // Response to a pending request (has 'id')
+    if ('id' in msg && msg.id !== undefined && ('result' in msg || 'error' in msg)) {
+      this.handleResponse(msg as { id: number; result?: unknown; error?: { code: number; message: string } });
+      return;
+    }
+
+    // Notification (has 'method', no 'id')
+    if ('method' in msg && !('id' in msg)) {
+      this.handleNotification(msg as { method: string; params?: Record<string, unknown> });
+      return;
+    }
+
+    // Request from agent (has 'method' AND 'id') — agent asking us for something
+    if ('method' in msg && 'id' in msg) {
+      this.handleAgentRequest(msg as { id: number; method: string; params?: Record<string, unknown> });
+      return;
+    }
+  }
+
+  private handleResponse(msg: { id: number; result?: unknown; error?: { code: number; message: string } }): void {
+    const pending = this.pendingRequests.get(msg.id);
+    if (!pending) return;
+
+    clearTimeout(pending.timer);
+    this.pendingRequests.delete(msg.id);
+
+    if (msg.error) {
+      pending.reject(new Error(`ACP error ${msg.error.code}: ${msg.error.message}`));
+    } else {
+      pending.resolve(msg.result);
+    }
+  }
+
+  private handleNotification(msg: { method: string; params?: Record<string, unknown> }): void {
+    if (msg.method !== 'session/update' || !msg.params) return;
+
+    const update = msg.params.update as Record<string, unknown> | undefined;
+    if (!update) return;
+
+    // Collect text from agent_message_chunk
+    if (update.sessionUpdate === 'agent_message_chunk') {
+      const content = update.content;
+      // Gemini ACP sends content as a single object {type, text} or an array [{type, text}]
+      const blocks: Array<{ type: string; text?: string }> = Array.isArray(content)
+        ? content as Array<{ type: string; text?: string }>
+        : content && typeof content === 'object'
+          ? [content as { type: string; text?: string }]
+          : [];
+      for (const block of blocks) {
+        if (block.type === 'text' && block.text) {
+          this.activePromptChunks.push(block.text);
+        }
+      }
+    }
+  }
+
+  /** Handle requests from the agent (e.g., session/request_permission). Reject them all. */
+  private handleAgentRequest(msg: { id: number; method: string; params?: Record<string, unknown> }): void {
+    if (!this.process?.stdin) return;
+
+    if (msg.method === 'session/request_permission') {
+      // Reject permission requests — we don't want tool use
+      const response = JSON.stringify({
+        jsonrpc: '2.0',
+        id: msg.id,
+        result: { outcome: { outcome: 'cancelled' } },
+      });
+      this.process.stdin.write(response + '\n');
+    } else {
+      // Unknown method — return error
+      const response = JSON.stringify({
+        jsonrpc: '2.0',
+        id: msg.id,
+        error: { code: -32601, message: 'Method not supported' },
+      });
+      this.process.stdin.write(response + '\n');
+    }
+  }
+
+  private sendRequest(method: string, params: Record<string, unknown>, timeoutMs: number): Promise<unknown> {
+    if (!this.process?.stdin) {
+      return Promise.reject(new Error('ACP process not started'));
+    }
+
+    const id = this.nextId++;
+
+    return new Promise((resolve, reject) => {
+      const timer = setTimeout(() => {
+        this.pendingRequests.delete(id);
+        // Kill the process on timeout — it's hung
+        this.cleanup();
+        reject(new Error(`ACP request '${method}' timed out after ${timeoutMs}ms`));
+      }, timeoutMs);
+
+      this.pendingRequests.set(id, { resolve, reject, timer });
+
+      const msg = JSON.stringify({ jsonrpc: '2.0', id, method, params });
+      this.process!.stdin!.write(msg + '\n');
+    });
+  }
+
+  private cleanup(): void {
+    this.ready = false;
+    this.initPromise = null;
+    this.sessionId = null;
+    this.activePromptChunks = [];
+
+    // Reject all pending requests
+    for (const [id, pending] of this.pendingRequests) {
+      clearTimeout(pending.timer);
+      pending.reject(new Error('ACP client disposed'));
+      this.pendingRequests.delete(id);
+    }
+
+    if (this.readline) {
+      this.readline.close();
+      this.readline = null;
+    }
+
+    if (this.process) {
+      this.process.kill('SIGTERM');
+      this.process = null;
+    }
+  }
+}

src/mcplocal/src/providers/gemini-acp.ts

@@ -0,0 +1,153 @@
+import { execFile } from 'node:child_process';
+import { promisify } from 'node:util';
+import type { LlmProvider, CompletionOptions, CompletionResult } from './types.js';
+import { AcpClient } from './acp-client.js';
+import type { AcpClientConfig } from './acp-client.js';
+
+const execFileAsync = promisify(execFile);
+
+export interface GeminiAcpConfig {
+  binaryPath?: string;
+  defaultModel?: string;
+  requestTimeoutMs?: number;
+  initTimeoutMs?: number;
+  /** Idle TTL for pooled sessions in ms (default: 8 hours) */
+  idleTtlMs?: number;
+  /** Override for testing — passed through to AcpClient */
+  spawn?: AcpClientConfig['spawn'];
+}
+
+interface PoolEntry {
+  client: AcpClient;
+  lastUsed: number;
+  queue: Promise<void>;
+}
+
+/**
+ * Gemini CLI provider using ACP (Agent Client Protocol) mode.
+ *
+ * Maintains a pool of persistent subprocesses keyed by model name.
+ * Each model gets its own `gemini --experimental-acp` subprocess with
+ * a serial request queue. Idle sessions are evicted after 8 hours.
+ *
+ * NOTE: Gemini ACP currently doesn't support per-session model selection,
+ * so all sessions use the same model. The pool infrastructure is ready for
+ * when vLLM/OpenAI providers are added (they support per-request model).
+ */
+export class GeminiAcpProvider implements LlmProvider {
+  readonly name = 'gemini-cli';
+  private pool = new Map<string, PoolEntry>();
+  private binaryPath: string;
+  private defaultModel: string;
+  private readonly requestTimeoutMs: number;
+  private readonly initTimeoutMs: number;
+  private readonly idleTtlMs: number;
+  private readonly spawnOverride?: AcpClientConfig['spawn'];
+
+  constructor(config?: GeminiAcpConfig) {
+    this.binaryPath = config?.binaryPath ?? 'gemini';
+    this.defaultModel = config?.defaultModel ?? 'gemini-2.5-flash';
+    this.requestTimeoutMs = config?.requestTimeoutMs ?? 60_000;
+    this.initTimeoutMs = config?.initTimeoutMs ?? 30_000;
+    this.idleTtlMs = config?.idleTtlMs ?? 8 * 60 * 60 * 1000; // 8 hours
+    if (config?.spawn) this.spawnOverride = config.spawn;
+  }
+
+  async complete(options: CompletionOptions): Promise<CompletionResult> {
+    const model = options.model ?? this.defaultModel;
+    const entry = this.getOrCreateEntry(model);
+    entry.lastUsed = Date.now();
+
+    this.evictIdle();
+
+    return this.enqueue(entry, () => this.doComplete(entry.client, options));
+  }
+
+  async listModels(): Promise<string[]> {
+    return ['gemini-2.5-flash', 'gemini-2.5-pro', 'gemini-2.0-flash'];
+  }
+
+  async isAvailable(): Promise<boolean> {
+    try {
+      await execFileAsync(this.binaryPath, ['--version'], { timeout: 5000 });
+      return true;
+    } catch {
+      return false;
+    }
+  }
+
+  dispose(): void {
+    for (const entry of this.pool.values()) {
+      entry.client.dispose();
+    }
+    this.pool.clear();
+  }
+
+  /** Number of active pool entries (for testing). */
+  get poolSize(): number {
+    return this.pool.size;
+  }
+
+  // --- Private ---
+
+  private getOrCreateEntry(model: string): PoolEntry {
+    const existing = this.pool.get(model);
+    if (existing) return existing;
+
+    const acpConfig: AcpClientConfig = {
+      binaryPath: this.binaryPath,
+      model,
+      requestTimeoutMs: this.requestTimeoutMs,
+      initTimeoutMs: this.initTimeoutMs,
+    };
+    if (this.spawnOverride) acpConfig.spawn = this.spawnOverride;
+
+    const entry: PoolEntry = {
+      client: new AcpClient(acpConfig),
+      lastUsed: Date.now(),
+      queue: Promise.resolve(),
+    };
+    this.pool.set(model, entry);
+    return entry;
+  }
+
+  private evictIdle(): void {
+    const now = Date.now();
+    for (const [model, entry] of this.pool) {
+      if (now - entry.lastUsed > this.idleTtlMs) {
+        entry.client.dispose();
+        this.pool.delete(model);
+      }
+    }
+  }
+
+  private async doComplete(client: AcpClient, options: CompletionOptions): Promise<CompletionResult> {
+    const prompt = options.messages
+      .map((m) => {
+        if (m.role === 'system') return `System: ${m.content}`;
+        if (m.role === 'user') return m.content;
+        if (m.role === 'assistant') return `Assistant: ${m.content}`;
+        return m.content;
+      })
+      .join('\n\n');
+
+    const content = await client.prompt(prompt);
+
+    return {
+      content: content.trim(),
+      toolCalls: [],
+      usage: { promptTokens: 0, completionTokens: 0, totalTokens: 0 },
+      finishReason: 'stop',
+    };
+  }
+
+  private enqueue<T>(entry: PoolEntry, fn: () => Promise<T>): Promise<T> {
+    const result = new Promise<T>((resolve, reject) => {
+      entry.queue = entry.queue.then(
+        () => fn().then(resolve, reject),
+        () => fn().then(resolve, reject),
+      );
+    });
+    return result;
+  }
+}

src/mcplocal/src/providers/index.ts

@@ -9,4 +9,8 @@ export { GeminiCliProvider } from './gemini-cli.js';
 export type { GeminiCliConfig } from './gemini-cli.js';
 export { DeepSeekProvider } from './deepseek.js';
 export type { DeepSeekConfig } from './deepseek.js';
+export { GeminiAcpProvider } from './gemini-acp.js';
+export type { GeminiAcpConfig } from './gemini-acp.js';
+export { AcpClient } from './acp-client.js';
+export type { AcpClientConfig } from './acp-client.js';
 export { ProviderRegistry } from './registry.js';

src/mcplocal/src/providers/registry.ts

@@ -45,4 +45,11 @@ export class ProviderRegistry {
   getActiveName(): string | null {
     return this.activeProvider;
   }
+
+  /** Dispose all registered providers that have a dispose method. */
+  disposeAll(): void {
+    for (const provider of this.providers.values()) {
+      provider.dispose?.();
+    }
+  }
 }

src/mcplocal/src/providers/types.ts

@@ -53,4 +53,6 @@ export interface LlmProvider {
   listModels(): Promise<string[]>;
   /** Check if the provider is configured and reachable */
   isAvailable(): Promise<boolean>;
+  /** Optional cleanup for providers with persistent resources (e.g., subprocesses). */
+  dispose?(): void;
 }

src/mcplocal/tests/acp-client.test.ts

@@ -0,0 +1,486 @@
+import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
+import { EventEmitter, Readable } from 'node:stream';
+import { AcpClient } from '../src/providers/acp-client.js';
+import type { AcpClientConfig } from '../src/providers/acp-client.js';
+
+/**
+ * Creates a mock child process that speaks ACP protocol.
+ * Returns the mock process and helpers to send responses.
+ */
+function createMockProcess() {
+  const stdin = {
+    write: vi.fn(),
+    writable: true,
+  };
+
+  const stdoutEmitter = new EventEmitter();
+  const stdout = Object.assign(stdoutEmitter, {
+    readable: true,
+    // readline needs these
+    [Symbol.asyncIterator]: undefined,
+    pause: vi.fn(),
+    resume: vi.fn(),
+    isPaused: () => false,
+    setEncoding: vi.fn(),
+    read: vi.fn(),
+    destroy: vi.fn(),
+    pipe: vi.fn(),
+    unpipe: vi.fn(),
+    unshift: vi.fn(),
+    wrap: vi.fn(),
+  }) as unknown as Readable;
+
+  const proc = Object.assign(new EventEmitter(), {
+    stdin,
+    stdout,
+    stderr: new EventEmitter(),
+    pid: 12345,
+    killed: false,
+    kill: vi.fn(function (this: { killed: boolean }) {
+      this.killed = true;
+    }),
+  });
+
+  /** Send a line of JSON from the "agent" to our client */
+  function sendLine(data: unknown) {
+    stdoutEmitter.emit('data', Buffer.from(JSON.stringify(data) + '\n'));
+  }
+
+  /** Send a JSON-RPC response */
+  function sendResponse(id: number, result: unknown) {
+    sendLine({ jsonrpc: '2.0', id, result });
+  }
+
+  /** Send a JSON-RPC error */
+  function sendError(id: number, code: number, message: string) {
+    sendLine({ jsonrpc: '2.0', id, error: { code, message } });
+  }
+
+  /** Send a session/update notification with agent_message_chunk */
+  function sendChunk(sessionId: string, text: string) {
+    sendLine({
+      jsonrpc: '2.0',
+      method: 'session/update',
+      params: {
+        sessionId,
+        update: {
+          sessionUpdate: 'agent_message_chunk',
+          content: [{ type: 'text', text }],
+        },
+      },
+    });
+  }
+
+  /** Send a session/request_permission request */
+  function sendPermissionRequest(id: number, sessionId: string) {
+    sendLine({
+      jsonrpc: '2.0',
+      id,
+      method: 'session/request_permission',
+      params: { sessionId },
+    });
+  }
+
+  return { proc, stdin, stdout: stdoutEmitter, sendLine, sendResponse, sendError, sendChunk, sendPermissionRequest };
+}
+
+function createConfig(overrides?: Partial<AcpClientConfig>): AcpClientConfig {
+  return {
+    binaryPath: '/usr/bin/gemini',
+    model: 'gemini-2.5-flash',
+    requestTimeoutMs: 5000,
+    initTimeoutMs: 5000,
+    ...overrides,
+  };
+}
+
+describe('AcpClient', () => {
+  let client: AcpClient;
+  let mock: ReturnType<typeof createMockProcess>;
+
+  beforeEach(() => {
+    mock = createMockProcess();
+  });
+
+  afterEach(() => {
+    client?.dispose();
+  });
+
+  function createClient(configOverrides?: Partial<AcpClientConfig>) {
+    const config = createConfig({
+      spawn: (() => mock.proc) as unknown as AcpClientConfig['spawn'],
+      ...configOverrides,
+    });
+    client = new AcpClient(config);
+    return client;
+  }
+
+  /** Helper: auto-respond to the initialize + session/new handshake */
+  function autoHandshake(sessionId = 'test-session') {
+    mock.stdin.write.mockImplementation((data: string) => {
+      const msg = JSON.parse(data.trim()) as { id: number; method: string };
+      if (msg.method === 'initialize') {
+        // Respond async to simulate real behavior
+        setImmediate(() => mock.sendResponse(msg.id, {
+          protocolVersion: 1,
+          agentInfo: { name: 'gemini-cli', version: '1.0.0' },
+        }));
+      } else if (msg.method === 'session/new') {
+        setImmediate(() => mock.sendResponse(msg.id, { sessionId }));
+      }
+    });
+  }
+
+  describe('ensureReady', () => {
+    it('spawns process and completes ACP handshake', async () => {
+      createClient();
+      autoHandshake();
+
+      await client.ensureReady();
+
+      expect(client.isAlive).toBe(true);
+      // Verify initialize was sent
+      const calls = mock.stdin.write.mock.calls.map((c) => JSON.parse(c[0] as string));
+      expect(calls[0].method).toBe('initialize');
+      expect(calls[0].params.protocolVersion).toBe(1);
+      expect(calls[0].params.clientInfo.name).toBe('mcpctl');
+      // Verify session/new was sent
+      expect(calls[1].method).toBe('session/new');
+      expect(calls[1].params.cwd).toBe('/tmp');
+      expect(calls[1].params.mcpServers).toEqual([]);
+    });
+
+    it('is idempotent when already ready', async () => {
+      createClient();
+      autoHandshake();
+
+      await client.ensureReady();
+      await client.ensureReady();
+
+      // Should only have sent initialize + session/new once
+      const calls = mock.stdin.write.mock.calls;
+      expect(calls.length).toBe(2);
+    });
+
+    it('shares init promise for concurrent calls', async () => {
+      createClient();
+      autoHandshake();
+
+      const p1 = client.ensureReady();
+      const p2 = client.ensureReady();
+
+      await Promise.all([p1, p2]);
+      expect(mock.stdin.write.mock.calls.length).toBe(2);
+    });
+  });
+
+  describe('prompt', () => {
+    it('sends session/prompt and collects agent_message_chunk text', async () => {
+      createClient();
+      const sessionId = 'sess-123';
+      autoHandshake(sessionId);
+
+      await client.ensureReady();
+
+      // Now set up the prompt response handler
+      mock.stdin.write.mockImplementation((data: string) => {
+        const msg = JSON.parse(data.trim()) as { id: number; method: string };
+        if (msg.method === 'session/prompt') {
+          setImmediate(() => {
+            mock.sendChunk(sessionId, 'Hello ');
+            mock.sendChunk(sessionId, 'world!');
+            mock.sendResponse(msg.id, { stopReason: 'end_turn' });
+          });
+        }
+      });
+
+      const result = await client.prompt('Say hello');
+      expect(result).toBe('Hello world!');
+    });
+
+    it('handles multi-block content in a single chunk', async () => {
+      createClient();
+      autoHandshake('sess-1');
+      await client.ensureReady();
+
+      mock.stdin.write.mockImplementation((data: string) => {
+        const msg = JSON.parse(data.trim()) as { id: number; method: string };
+        if (msg.method === 'session/prompt') {
+          setImmediate(() => {
+            mock.sendLine({
+              jsonrpc: '2.0',
+              method: 'session/update',
+              params: {
+                sessionId: 'sess-1',
+                update: {
+                  sessionUpdate: 'agent_message_chunk',
+                  content: [
+                    { type: 'text', text: 'Part A' },
+                    { type: 'text', text: ' Part B' },
+                  ],
+                },
+              },
+            });
+            mock.sendResponse(msg.id, { stopReason: 'end_turn' });
+          });
+        }
+      });
+
+      const result = await client.prompt('test');
+      expect(result).toBe('Part A Part B');
+    });
+
+    it('handles single-object content (real Gemini ACP format)', async () => {
+      createClient();
+      autoHandshake('sess-1');
+      await client.ensureReady();
+
+      mock.stdin.write.mockImplementation((data: string) => {
+        const msg = JSON.parse(data.trim()) as { id: number; method: string };
+        if (msg.method === 'session/prompt') {
+          setImmediate(() => {
+            // Real Gemini ACP sends content as a single object, not an array
+            mock.sendLine({
+              jsonrpc: '2.0',
+              method: 'session/update',
+              params: {
+                sessionId: 'sess-1',
+                update: {
+                  sessionUpdate: 'agent_message_chunk',
+                  content: { type: 'text', text: 'ok' },
+                },
+              },
+            });
+            mock.sendResponse(msg.id, { stopReason: 'end_turn' });
+          });
+        }
+      });
+
+      const result = await client.prompt('test');
+      expect(result).toBe('ok');
+    });
+
+    it('ignores agent_thought_chunk notifications', async () => {
+      createClient();
+      autoHandshake('sess-1');
+      await client.ensureReady();
+
+      mock.stdin.write.mockImplementation((data: string) => {
+        const msg = JSON.parse(data.trim()) as { id: number; method: string };
+        if (msg.method === 'session/prompt') {
+          setImmediate(() => {
+            // Gemini sends thought chunks before message chunks
+            mock.sendLine({
+              jsonrpc: '2.0',
+              method: 'session/update',
+              params: {
+                sessionId: 'sess-1',
+                update: {
+                  sessionUpdate: 'agent_thought_chunk',
+                  content: { type: 'text', text: 'Thinking about it...' },
+                },
+              },
+            });
+            mock.sendLine({
+              jsonrpc: '2.0',
+              method: 'session/update',
+              params: {
+                sessionId: 'sess-1',
+                update: {
+                  sessionUpdate: 'agent_message_chunk',
+                  content: { type: 'text', text: 'ok' },
+                },
+              },
+            });
+            mock.sendResponse(msg.id, { stopReason: 'end_turn' });
+          });
+        }
+      });
+
+      const result = await client.prompt('test');
+      expect(result).toBe('ok');
+    });
+
+    it('calls ensureReady automatically (lazy init)', async () => {
+      createClient();
+      autoHandshake('sess-auto');
+
+      // After handshake, handle prompts
+      const originalWrite = mock.stdin.write;
+      let handshakeDone = false;
+      mock.stdin.write.mockImplementation((data: string) => {
+        const msg = JSON.parse(data.trim()) as { id: number; method: string };
+        if (msg.method === 'initialize') {
+          setImmediate(() => mock.sendResponse(msg.id, { protocolVersion: 1 }));
+        } else if (msg.method === 'session/new') {
+          setImmediate(() => {
+            mock.sendResponse(msg.id, { sessionId: 'sess-auto' });
+            handshakeDone = true;
+          });
+        } else if (msg.method === 'session/prompt') {
+          setImmediate(() => {
+            mock.sendChunk('sess-auto', 'ok');
+            mock.sendResponse(msg.id, { stopReason: 'end_turn' });
+          });
+        }
+      });
+
+      // Call prompt directly without ensureReady
+      const result = await client.prompt('test');
+      expect(result).toBe('ok');
+    });
+  });
+
+  describe('auto-restart', () => {
+    it('restarts after process exit', async () => {
+      createClient();
+      autoHandshake('sess-1');
+      await client.ensureReady();
+      expect(client.isAlive).toBe(true);
+
+      // Simulate process exit
+      mock.proc.killed = true;
+      mock.proc.emit('exit', 1);
+      expect(client.isAlive).toBe(false);
+
+      // Create a new mock for the respawned process
+      mock = createMockProcess();
+      // Update the spawn function to return new mock
+      (client as unknown as { config: { spawn: unknown } }).config.spawn = () => mock.proc;
+      autoHandshake('sess-2');
+
+      await client.ensureReady();
+      expect(client.isAlive).toBe(true);
+    });
+  });
+
+  describe('timeout', () => {
+    it('kills process and rejects on request timeout', async () => {
+      createClient({ requestTimeoutMs: 50 });
+      autoHandshake('sess-1');
+      await client.ensureReady();
+
+      // Don't respond to the prompt — let it timeout
+      mock.stdin.write.mockImplementation(() => {});
+
+      await expect(client.prompt('test')).rejects.toThrow('timed out');
+      expect(client.isAlive).toBe(false);
+    });
+
+    it('rejects on init timeout', async () => {
+      createClient({ initTimeoutMs: 50 });
+      // Don't respond to initialize
+      mock.stdin.write.mockImplementation(() => {});
+
+      await expect(client.ensureReady()).rejects.toThrow('timed out');
+    });
+  });
+
+  describe('error handling', () => {
+    it('rejects on ACP error response', async () => {
+      createClient();
+      mock.stdin.write.mockImplementation((data: string) => {
+        const msg = JSON.parse(data.trim()) as { id: number; method: string };
+        setImmediate(() => mock.sendError(msg.id, -32603, 'Internal error'));
+      });
+
+      await expect(client.ensureReady()).rejects.toThrow('ACP error -32603: Internal error');
+    });
+
+    it('rejects pending requests on process crash', async () => {
+      createClient();
+      autoHandshake('sess-1');
+      await client.ensureReady();
+
+      // Override write so prompt sends but gets no response; then crash the process
+      mock.stdin.write.mockImplementation(() => {
+        // After the prompt is sent, simulate a process crash
+        setImmediate(() => {
+          mock.proc.killed = true;
+          mock.proc.emit('exit', 1);
+        });
+      });
+
+      const promptPromise = client.prompt('test');
+      await expect(promptPromise).rejects.toThrow('process exited');
+    });
+  });
+
+  describe('permission requests', () => {
+    it('rejects session/request_permission from agent', async () => {
+      createClient();
+      autoHandshake('sess-1');
+      await client.ensureReady();
+
+      mock.stdin.write.mockImplementation((data: string) => {
+        const msg = JSON.parse(data.trim()) as { id: number; method: string };
+        if (msg.method === 'session/prompt') {
+          setImmediate(() => {
+            // Agent asks for permission first
+            mock.sendPermissionRequest(100, 'sess-1');
+            // Then provides the actual response
+            mock.sendChunk('sess-1', 'done');
+            mock.sendResponse(msg.id, { stopReason: 'end_turn' });
+          });
+        }
+      });
+
+      const result = await client.prompt('test');
+      expect(result).toBe('done');
+
+      // Verify we sent a rejection for the permission request
+      const writes = mock.stdin.write.mock.calls.map((c) => {
+        try { return JSON.parse(c[0] as string); } catch { return null; }
+      }).filter(Boolean);
+      const rejection = writes.find((w: Record<string, unknown>) => w.id === 100);
+      expect(rejection).toBeTruthy();
+      expect((rejection as { result: { outcome: { outcome: string } } }).result.outcome.outcome).toBe('cancelled');
+    });
+  });
+
+  describe('dispose', () => {
+    it('kills process and rejects pending', async () => {
+      createClient();
+      autoHandshake('sess-1');
+      await client.ensureReady();
+
+      // Override write so prompt is sent but gets no response; then dispose
+      mock.stdin.write.mockImplementation(() => {
+        setImmediate(() => client.dispose());
+      });
+
+      const promptPromise = client.prompt('test');
+      await expect(promptPromise).rejects.toThrow('disposed');
+      expect(mock.proc.kill).toHaveBeenCalledWith('SIGTERM');
+    });
+
+    it('is safe to call multiple times', () => {
+      createClient();
+      client.dispose();
+      client.dispose();
+      // No error thrown
+    });
+  });
+
+  describe('isAlive', () => {
+    it('returns false before init', () => {
+      createClient();
+      expect(client.isAlive).toBe(false);
+    });
+
+    it('returns true after successful init', async () => {
+      createClient();
+      autoHandshake();
+      await client.ensureReady();
+      expect(client.isAlive).toBe(true);
+    });
+
+    it('returns false after dispose', async () => {
+      createClient();
+      autoHandshake();
+      await client.ensureReady();
+      client.dispose();
+      expect(client.isAlive).toBe(false);
+    });
+  });
+});

src/mcplocal/tests/gemini-acp.test.ts

@@ -0,0 +1,200 @@
+import { describe, it, expect, vi, beforeEach } from 'vitest';
+
+const mockEnsureReady = vi.fn(async () => {});
+const mockPrompt = vi.fn(async () => 'mock response');
+const mockDispose = vi.fn();
+
+vi.mock('../src/providers/acp-client.js', () => ({
+  AcpClient: vi.fn(function (this: Record<string, unknown>) {
+    this.ensureReady = mockEnsureReady;
+    this.prompt = mockPrompt;
+    this.dispose = mockDispose;
+  }),
+}));
+
+// Must import after mock setup
+const { GeminiAcpProvider } = await import('../src/providers/gemini-acp.js');
+
+describe('GeminiAcpProvider', () => {
+  let provider: InstanceType<typeof GeminiAcpProvider>;
+
+  beforeEach(() => {
+    vi.clearAllMocks();
+    mockPrompt.mockResolvedValue('mock response');
+    provider = new GeminiAcpProvider({ binaryPath: '/usr/bin/gemini', defaultModel: 'gemini-2.5-flash' });
+  });
+
+  describe('complete', () => {
+    it('builds prompt from messages and returns CompletionResult', async () => {
+      mockPrompt.mockResolvedValueOnce('The answer is 42.');
+
+      const result = await provider.complete({
+        messages: [
+          { role: 'system', content: 'You are helpful.' },
+          { role: 'user', content: 'What is the answer?' },
+        ],
+      });
+
+      expect(result.content).toBe('The answer is 42.');
+      expect(result.toolCalls).toEqual([]);
+      expect(result.finishReason).toBe('stop');
+
+      const promptText = mockPrompt.mock.calls[0][0] as string;
+      expect(promptText).toContain('System: You are helpful.');
+      expect(promptText).toContain('What is the answer?');
+    });
+
+    it('formats assistant messages with prefix', async () => {
+      mockPrompt.mockResolvedValueOnce('ok');
+
+      await provider.complete({
+        messages: [
+          { role: 'user', content: 'Hello' },
+          { role: 'assistant', content: 'Hi there' },
+          { role: 'user', content: 'How are you?' },
+        ],
+      });
+
+      const promptText = mockPrompt.mock.calls[0][0] as string;
+      expect(promptText).toContain('Assistant: Hi there');
+    });
+
+    it('trims response content', async () => {
+      mockPrompt.mockResolvedValueOnce('  padded response  \n');
+
+      const result = await provider.complete({
+        messages: [{ role: 'user', content: 'test' }],
+      });
+
+      expect(result.content).toBe('padded response');
+    });
+
+    it('serializes concurrent calls to same model', async () => {
+      const callOrder: number[] = [];
+      let callCount = 0;
+
+      mockPrompt.mockImplementation(async () => {
+        const myCall = ++callCount;
+        callOrder.push(myCall);
+        await new Promise((r) => setTimeout(r, 10));
+        return `response-${myCall}`;
+      });
+
+      const [r1, r2, r3] = await Promise.all([
+        provider.complete({ messages: [{ role: 'user', content: 'a' }] }),
+        provider.complete({ messages: [{ role: 'user', content: 'b' }] }),
+        provider.complete({ messages: [{ role: 'user', content: 'c' }] }),
+      ]);
+
+      expect(r1.content).toBe('response-1');
+      expect(r2.content).toBe('response-2');
+      expect(r3.content).toBe('response-3');
+      expect(callOrder).toEqual([1, 2, 3]);
+    });
+
+    it('continues queue after error', async () => {
+      mockPrompt
+        .mockRejectedValueOnce(new Error('first fails'))
+        .mockResolvedValueOnce('second works');
+
+      const results = await Promise.allSettled([
+        provider.complete({ messages: [{ role: 'user', content: 'a' }] }),
+        provider.complete({ messages: [{ role: 'user', content: 'b' }] }),
+      ]);
+
+      expect(results[0].status).toBe('rejected');
+      expect(results[1].status).toBe('fulfilled');
+      if (results[1].status === 'fulfilled') {
+        expect(results[1].value.content).toBe('second works');
+      }
+    });
+  });
+
+  describe('session pool', () => {
+    it('creates separate pool entries for different models', async () => {
+      mockPrompt.mockResolvedValue('ok');
+
+      await provider.complete({ messages: [{ role: 'user', content: 'a' }], model: 'gemini-2.5-flash' });
+      await provider.complete({ messages: [{ role: 'user', content: 'b' }], model: 'gemini-2.5-pro' });
+
+      expect(provider.poolSize).toBe(2);
+    });
+
+    it('reuses existing pool entry for same model', async () => {
+      mockPrompt.mockResolvedValue('ok');
+
+      await provider.complete({ messages: [{ role: 'user', content: 'a' }], model: 'gemini-2.5-flash' });
+      await provider.complete({ messages: [{ role: 'user', content: 'b' }], model: 'gemini-2.5-flash' });
+
+      expect(provider.poolSize).toBe(1);
+    });
+
+    it('uses defaultModel when no model specified', async () => {
+      mockPrompt.mockResolvedValue('ok');
+
+      await provider.complete({ messages: [{ role: 'user', content: 'a' }] });
+
+      expect(provider.poolSize).toBe(1);
+    });
+
+    it('evicts idle sessions', async () => {
+      // Use a very short TTL for testing
+      const shortTtl = new GeminiAcpProvider({
+        binaryPath: '/usr/bin/gemini',
+        defaultModel: 'gemini-2.5-flash',
+        idleTtlMs: 1, // 1ms TTL
+      });
+
+      mockPrompt.mockResolvedValue('ok');
+      await shortTtl.complete({ messages: [{ role: 'user', content: 'a' }], model: 'model-a' });
+      expect(shortTtl.poolSize).toBe(1);
+
+      // Wait for TTL to expire
+      await new Promise((r) => setTimeout(r, 10));
+
+      // Next complete call triggers eviction of old entry and creates new one
+      await shortTtl.complete({ messages: [{ role: 'user', content: 'b' }], model: 'model-b' });
+      // model-a should have been evicted, only model-b remains
+      expect(shortTtl.poolSize).toBe(1);
+      expect(mockDispose).toHaveBeenCalled();
+
+      shortTtl.dispose();
+    });
+
+    it('dispose kills all pooled clients', async () => {
+      mockPrompt.mockResolvedValue('ok');
+
+      await provider.complete({ messages: [{ role: 'user', content: 'a' }], model: 'model-a' });
+      await provider.complete({ messages: [{ role: 'user', content: 'b' }], model: 'model-b' });
+      expect(provider.poolSize).toBe(2);
+
+      provider.dispose();
+      expect(provider.poolSize).toBe(0);
+      expect(mockDispose).toHaveBeenCalledTimes(2);
+    });
+  });
+
+  describe('listModels', () => {
+    it('returns static model list', async () => {
+      const models = await provider.listModels();
+      expect(models).toContain('gemini-2.5-flash');
+      expect(models).toContain('gemini-2.5-pro');
+      expect(models).toContain('gemini-2.0-flash');
+    });
+  });
+
+  describe('dispose', () => {
+    it('delegates to all pooled AcpClients', async () => {
+      mockPrompt.mockResolvedValue('ok');
+      await provider.complete({ messages: [{ role: 'user', content: 'test' }] });
+      provider.dispose();
+      expect(mockDispose).toHaveBeenCalled();
+    });
+  });
+
+  describe('name', () => {
+    it('is gemini-cli for config compatibility', () => {
+      expect(provider.name).toBe('gemini-cli');
+    });
+  });
+});

src/mcplocal/tests/http/config.test.ts

@@ -1,5 +1,5 @@
-import { describe, it, expect, vi, afterEach } from 'vitest';
-import { loadLlmConfig } from '../../src/http/config.js';
+import { describe, it, expect, vi, afterEach, beforeEach } from 'vitest';
+import { loadLlmConfig, resetConfigCache } from '../../src/http/config.js';
 import { existsSync, readFileSync } from 'node:fs';
 
 vi.mock('node:fs', async () => {
@@ -11,6 +11,10 @@ vi.mock('node:fs', async () => {
   };
 });
 
+beforeEach(() => {
+  resetConfigCache();
+});
+
 afterEach(() => {
   vi.restoreAllMocks();
 });

src/mcplocal/tests/integration/e2e-flow.test.ts

@@ -6,13 +6,14 @@
  * (node:http) and a mock LLM provider. No Docker or external services needed.
  */
 
-import { describe, it, expect, beforeEach, afterEach, afterAll } from 'vitest';
+import { describe, it, expect, vi, beforeEach, afterEach, afterAll } from 'vitest';
 import { createServer, type Server, type IncomingMessage, type ServerResponse } from 'node:http';
 
 import { McpRouter } from '../../src/router.js';
 import { McpdUpstream } from '../../src/upstream/mcpd.js';
 import { McpdClient } from '../../src/http/mcpd-client.js';
 import { LlmProcessor, DEFAULT_PROCESSOR_CONFIG } from '../../src/llm/processor.js';
+import { ResponsePaginator } from '../../src/llm/pagination.js';
 import { ProviderRegistry } from '../../src/providers/registry.js';
 import { TieredHealthMonitor } from '../../src/health/tiered.js';
 import { refreshUpstreams } from '../../src/discovery.js';
@@ -1096,4 +1097,429 @@ describe('End-to-end integration: 3-tier architecture', () => {
       }
     });
   });
+
+  // -----------------------------------------------------------------------
+  // 8. Smart pagination through the full pipeline
+  // -----------------------------------------------------------------------
+  describe('Smart pagination', () => {
+    // Helper: generate a large JSON response (~100KB)
+    function makeLargeToolResult(): { flows: Array<{ id: string; type: string; label: string; wires: string[] }> } {
+      return {
+        flows: Array.from({ length: 200 }, (_, i) => ({
+          id: `flow-${String(i).padStart(4, '0')}`,
+          type: i % 3 === 0 ? 'function' : i % 3 === 1 ? 'http request' : 'inject',
+          label: `Node ${String(i)}: ${i % 3 === 0 ? 'Data transform' : i % 3 === 1 ? 'API call' : 'Timer trigger'}`,
+          wires: [`flow-${String(i + 1).padStart(4, '0')}`],
+        })),
+      };
+    }
+
+    it('paginates large tool response with smart AI summaries through router', async () => {
+      const largeResult = makeLargeToolResult();
+
+      mockMcpd = await startMockMcpd({
+        servers: [{ id: 'srv-nodered', name: 'node-red', transport: 'stdio' }],
+        proxyResponses: new Map([
+          ['srv-nodered:tools/list', {
+            result: { tools: [{ name: 'get_flows', description: 'Get all flows' }] },
+          }],
+          ['srv-nodered:tools/call', {
+            result: largeResult,
+          }],
+        ]),
+      });
+
+      const client = new McpdClient(mockMcpd.baseUrl, mockMcpd.config.expectedToken);
+      router = new McpRouter();
+      await refreshUpstreams(router, client);
+      await router.discoverTools();
+
+      // Set up paginator with LLM provider for smart summaries
+      const registry = new ProviderRegistry();
+      const completeFn = vi.fn().mockImplementation(() => ({
+        content: JSON.stringify([
+          { page: 1, summary: 'Function nodes and data transforms (flow-0000 through flow-0050)' },
+          { page: 2, summary: 'HTTP request nodes and API integrations (flow-0051 through flow-0100)' },
+          { page: 3, summary: 'Inject/timer nodes and triggers (flow-0101 through flow-0150)' },
+          { page: 4, summary: 'Remaining nodes and wire connections (flow-0151 through flow-0199)' },
+        ]),
+      }));
+      const mockProvider: LlmProvider = {
+        name: 'test-paginator',
+        isAvailable: () => true,
+        complete: completeFn,
+      };
+      registry.register(mockProvider);
+
+      // Low threshold so our response triggers pagination
+      const paginator = new ResponsePaginator(registry, {
+        sizeThreshold: 1000,
+        pageSize: 8000,
+      });
+      router.setPaginator(paginator);
+
+      // Call the tool — should get pagination index, not raw data
+      const response = await router.route({
+        jsonrpc: '2.0',
+        id: 'paginate-1',
+        method: 'tools/call',
+        params: { name: 'node-red/get_flows', arguments: {} },
+      });
+
+      expect(response.error).toBeUndefined();
+      const result = response.result as { content: Array<{ type: string; text: string }> };
+      expect(result.content).toHaveLength(1);
+      const indexText = result.content[0]!.text;
+
+      // Verify smart index with AI summaries
+      expect(indexText).toContain('AI-generated summaries');
+      expect(indexText).toContain('Function nodes and data transforms');
+      expect(indexText).toContain('HTTP request nodes');
+      expect(indexText).toContain('_resultId');
+      expect(indexText).toContain('_page');
+
+      // LLM was called to generate summaries
+      expect(completeFn).toHaveBeenCalledOnce();
+      const llmCall = completeFn.mock.calls[0]![0]!;
+      expect(llmCall.messages[0].role).toBe('system');
+      expect(llmCall.messages[1].content).toContain('node-red/get_flows');
+    });
+
+    it('retrieves specific pages after pagination via _resultId/_page', async () => {
+      const largeResult = makeLargeToolResult();
+
+      mockMcpd = await startMockMcpd({
+        servers: [{ id: 'srv-nodered', name: 'node-red', transport: 'stdio' }],
+        proxyResponses: new Map([
+          ['srv-nodered:tools/list', {
+            result: { tools: [{ name: 'get_flows', description: 'Get all flows' }] },
+          }],
+          ['srv-nodered:tools/call', {
+            result: largeResult,
+          }],
+        ]),
+      });
+
+      const client = new McpdClient(mockMcpd.baseUrl, mockMcpd.config.expectedToken);
+      router = new McpRouter();
+      await refreshUpstreams(router, client);
+      await router.discoverTools();
+
+      // Simple paginator (no LLM) for predictable behavior
+      const paginator = new ResponsePaginator(null, {
+        sizeThreshold: 1000,
+        pageSize: 8000,
+      });
+      router.setPaginator(paginator);
+
+      // First call — get the pagination index
+      const indexResponse = await router.route({
+        jsonrpc: '2.0',
+        id: 'idx-1',
+        method: 'tools/call',
+        params: { name: 'node-red/get_flows', arguments: {} },
+      });
+
+      expect(indexResponse.error).toBeUndefined();
+      const indexResult = indexResponse.result as { content: Array<{ text: string }> };
+      const indexText = indexResult.content[0]!.text;
+      const resultIdMatch = /"_resultId": "([^"]+)"/.exec(indexText);
+      expect(resultIdMatch).not.toBeNull();
+      const resultId = resultIdMatch![1]!;
+
+      // Second call — retrieve page 1 via _resultId/_page
+      const page1Response = await router.route({
+        jsonrpc: '2.0',
+        id: 'page-1',
+        method: 'tools/call',
+        params: {
+          name: 'node-red/get_flows',
+          arguments: { _resultId: resultId, _page: 1 },
+        },
+      });
+
+      expect(page1Response.error).toBeUndefined();
+      const page1Result = page1Response.result as { content: Array<{ text: string }> };
+      expect(page1Result.content[0]!.text).toContain('Page 1/');
+      // Page content should contain flow data
+      expect(page1Result.content[0]!.text).toContain('flow-');
+
+      // Third call — retrieve page 2
+      const page2Response = await router.route({
+        jsonrpc: '2.0',
+        id: 'page-2',
+        method: 'tools/call',
+        params: {
+          name: 'node-red/get_flows',
+          arguments: { _resultId: resultId, _page: 2 },
+        },
+      });
+
+      expect(page2Response.error).toBeUndefined();
+      const page2Result = page2Response.result as { content: Array<{ text: string }> };
+      expect(page2Result.content[0]!.text).toContain('Page 2/');
+    });
+
+    it('retrieves full content with _page=all', async () => {
+      const largeResult = makeLargeToolResult();
+
+      mockMcpd = await startMockMcpd({
+        servers: [{ id: 'srv-nodered', name: 'node-red', transport: 'stdio' }],
+        proxyResponses: new Map([
+          ['srv-nodered:tools/list', {
+            result: { tools: [{ name: 'get_flows', description: 'Get all flows' }] },
+          }],
+          ['srv-nodered:tools/call', {
+            result: largeResult,
+          }],
+        ]),
+      });
+
+      const client = new McpdClient(mockMcpd.baseUrl, mockMcpd.config.expectedToken);
+      router = new McpRouter();
+      await refreshUpstreams(router, client);
+      await router.discoverTools();
+
+      const paginator = new ResponsePaginator(null, {
+        sizeThreshold: 1000,
+        pageSize: 8000,
+      });
+      router.setPaginator(paginator);
+
+      // Get index
+      const indexResponse = await router.route({
+        jsonrpc: '2.0',
+        id: 'all-idx',
+        method: 'tools/call',
+        params: { name: 'node-red/get_flows', arguments: {} },
+      });
+      const indexText = (indexResponse.result as { content: Array<{ text: string }> }).content[0]!.text;
+      const resultId = /"_resultId": "([^"]+)"/.exec(indexText)![1]!;
+
+      // Request all pages
+      const allResponse = await router.route({
+        jsonrpc: '2.0',
+        id: 'all-1',
+        method: 'tools/call',
+        params: {
+          name: 'node-red/get_flows',
+          arguments: { _resultId: resultId, _page: 'all' },
+        },
+      });
+
+      expect(allResponse.error).toBeUndefined();
+      const allResult = allResponse.result as { content: Array<{ text: string }> };
+      // Full response should contain the original JSON
+      const fullText = allResult.content[0]!.text;
+      expect(fullText).toContain('flow-0000');
+      expect(fullText).toContain('flow-0199');
+      // Should be the full serialized result
+      expect(JSON.parse(fullText)).toEqual(largeResult);
+    });
+
+    it('falls back to simple index when LLM fails', async () => {
+      const largeResult = makeLargeToolResult();
+
+      mockMcpd = await startMockMcpd({
+        servers: [{ id: 'srv-nodered', name: 'node-red', transport: 'stdio' }],
+        proxyResponses: new Map([
+          ['srv-nodered:tools/list', {
+            result: { tools: [{ name: 'get_flows', description: 'Get all flows' }] },
+          }],
+          ['srv-nodered:tools/call', {
+            result: largeResult,
+          }],
+        ]),
+      });
+
+      const client = new McpdClient(mockMcpd.baseUrl, mockMcpd.config.expectedToken);
+      router = new McpRouter();
+      await refreshUpstreams(router, client);
+      await router.discoverTools();
+
+      // Set up paginator with a failing LLM
+      const registry = new ProviderRegistry();
+      registry.register(createFailingLlmProvider('broken-llm'));
+      const paginator = new ResponsePaginator(registry, {
+        sizeThreshold: 1000,
+        pageSize: 8000,
+      });
+      router.setPaginator(paginator);
+
+      const response = await router.route({
+        jsonrpc: '2.0',
+        id: 'fallback-idx',
+        method: 'tools/call',
+        params: { name: 'node-red/get_flows', arguments: {} },
+      });
+
+      expect(response.error).toBeUndefined();
+      const text = (response.result as { content: Array<{ text: string }> }).content[0]!.text;
+      // Should still paginate, just without AI summaries
+      expect(text).toContain('_resultId');
+      expect(text).not.toContain('AI-generated summaries');
+      expect(text).toContain('Page 1:');
+    });
+
+    it('returns expired cache message for stale _resultId', async () => {
+      router = new McpRouter();
+      const paginator = new ResponsePaginator(null, { sizeThreshold: 100, pageSize: 50 });
+      router.setPaginator(paginator);
+
+      // Try to retrieve a page with an unknown resultId
+      const response = await router.route({
+        jsonrpc: '2.0',
+        id: 'stale-1',
+        method: 'tools/call',
+        params: {
+          name: 'anything/tool',
+          arguments: { _resultId: 'nonexistent-id', _page: 1 },
+        },
+      });
+
+      expect(response.error).toBeUndefined();
+      const text = (response.result as { content: Array<{ text: string }> }).content[0]!.text;
+      expect(text).toContain('expired');
+      expect(text).toContain('re-call');
+    });
+
+    it('skips pagination for small responses', async () => {
+      mockMcpd = await startMockMcpd({
+        servers: [{ id: 'srv-small', name: 'smallserver', transport: 'stdio' }],
+        proxyResponses: new Map([
+          ['srv-small:tools/list', {
+            result: { tools: [{ name: 'get_status', description: 'Get status' }] },
+          }],
+          ['srv-small:tools/call', {
+            result: { status: 'ok', uptime: 12345 },
+          }],
+        ]),
+      });
+
+      const client = new McpdClient(mockMcpd.baseUrl, mockMcpd.config.expectedToken);
+      router = new McpRouter();
+      await refreshUpstreams(router, client);
+      await router.discoverTools();
+
+      const paginator = new ResponsePaginator(null, { sizeThreshold: 80000, pageSize: 40000 });
+      router.setPaginator(paginator);
+
+      const response = await router.route({
+        jsonrpc: '2.0',
+        id: 'small-1',
+        method: 'tools/call',
+        params: { name: 'smallserver/get_status', arguments: {} },
+      });
+
+      expect(response.error).toBeUndefined();
+      // Should return the raw result directly, not a pagination index
+      expect(response.result).toEqual({ status: 'ok', uptime: 12345 });
+    });
+
+    it('handles markdown-fenced LLM responses (Gemini quirk)', async () => {
+      const largeResult = makeLargeToolResult();
+
+      mockMcpd = await startMockMcpd({
+        servers: [{ id: 'srv-nodered', name: 'node-red', transport: 'stdio' }],
+        proxyResponses: new Map([
+          ['srv-nodered:tools/list', {
+            result: { tools: [{ name: 'get_flows', description: 'Get all flows' }] },
+          }],
+          ['srv-nodered:tools/call', {
+            result: largeResult,
+          }],
+        ]),
+      });
+
+      const client = new McpdClient(mockMcpd.baseUrl, mockMcpd.config.expectedToken);
+      router = new McpRouter();
+      await refreshUpstreams(router, client);
+      await router.discoverTools();
+
+      // Simulate Gemini wrapping JSON in ```json fences
+      const registry = new ProviderRegistry();
+      const mockProvider: LlmProvider = {
+        name: 'gemini-mock',
+        isAvailable: () => true,
+        complete: vi.fn().mockResolvedValue({
+          content: '```json\n' + JSON.stringify([
+            { page: 1, summary: 'Climate automation flows' },
+            { page: 2, summary: 'Lighting control flows' },
+          ]) + '\n```',
+        }),
+      };
+      registry.register(mockProvider);
+
+      const paginator = new ResponsePaginator(registry, {
+        sizeThreshold: 1000,
+        pageSize: 8000,
+      });
+      router.setPaginator(paginator);
+
+      const response = await router.route({
+        jsonrpc: '2.0',
+        id: 'fence-1',
+        method: 'tools/call',
+        params: { name: 'node-red/get_flows', arguments: {} },
+      });
+
+      expect(response.error).toBeUndefined();
+      const text = (response.result as { content: Array<{ text: string }> }).content[0]!.text;
+      // Fences were stripped — smart summaries should appear
+      expect(text).toContain('AI-generated summaries');
+      expect(text).toContain('Climate automation flows');
+      expect(text).toContain('Lighting control flows');
+    });
+
+    it('passes model override to LLM when project has custom model', async () => {
+      const largeResult = makeLargeToolResult();
+
+      mockMcpd = await startMockMcpd({
+        servers: [{ id: 'srv-nodered', name: 'node-red', transport: 'stdio' }],
+        proxyResponses: new Map([
+          ['srv-nodered:tools/list', {
+            result: { tools: [{ name: 'get_flows', description: 'Get all flows' }] },
+          }],
+          ['srv-nodered:tools/call', {
+            result: largeResult,
+          }],
+        ]),
+      });
+
+      const client = new McpdClient(mockMcpd.baseUrl, mockMcpd.config.expectedToken);
+      router = new McpRouter();
+      await refreshUpstreams(router, client);
+      await router.discoverTools();
+
+      const registry = new ProviderRegistry();
+      const completeFn = vi.fn().mockResolvedValue({
+        content: JSON.stringify([{ page: 1, summary: 'test' }]),
+      });
+      const mockProvider: LlmProvider = {
+        name: 'test-model-override',
+        isAvailable: () => true,
+        complete: completeFn,
+      };
+      registry.register(mockProvider);
+
+      // Paginator with per-project model override
+      const paginator = new ResponsePaginator(registry, {
+        sizeThreshold: 1000,
+        pageSize: 80000, // One big page so we get exactly 1 summary
+      }, 'gemini-2.5-pro');
+      router.setPaginator(paginator);
+
+      await router.route({
+        jsonrpc: '2.0',
+        id: 'model-1',
+        method: 'tools/call',
+        params: { name: 'node-red/get_flows', arguments: {} },
+      });
+
+      // Verify the model was passed through to the LLM call
+      expect(completeFn).toHaveBeenCalledOnce();
+      const llmOpts = completeFn.mock.calls[0]![0]!;
+      expect(llmOpts.model).toBe('gemini-2.5-pro');
+    });
+  });
 });

src/mcplocal/tests/llm-config.test.ts

@@ -25,7 +25,7 @@ describe('createProviderFromConfig', () => {
     expect(registry.getActive()).toBeNull();
   });
 
-  it('creates gemini-cli provider', async () => {
+  it('creates gemini-cli provider using ACP', async () => {
     const store = mockSecretStore();
     const registry = await createProviderFromConfig(
       { provider: 'gemini-cli', model: 'gemini-2.5-flash', binaryPath: '/usr/bin/gemini' },
@@ -33,6 +33,8 @@ describe('createProviderFromConfig', () => {
     );
     expect(registry.getActive()).not.toBeNull();
     expect(registry.getActive()!.name).toBe('gemini-cli');
+    // ACP provider has dispose method
+    expect(typeof registry.getActive()!.dispose).toBe('function');
   });
 
   it('creates ollama provider', async () => {

src/mcplocal/tests/pagination.test.ts

@@ -150,6 +150,25 @@ describe('ResponsePaginator', () => {
       expect(text).toContain('HTTP request nodes and API integrations');
     });
 
+    it('strips markdown code fences from LLM JSON response', async () => {
+      const summaries = [
+        { page: 1, summary: 'Config section' },
+        { page: 2, summary: 'Data section' },
+      ];
+      // Gemini often wraps JSON in ```json ... ``` fences
+      const fenced = '```json\n' + JSON.stringify(summaries) + '\n```';
+      const registry = makeProvider(fenced);
+      const paginator = new ResponsePaginator(registry, { sizeThreshold: 100, pageSize: 60 });
+      const raw = makeLargeStringWithNewlines(150);
+      const result = await paginator.paginate('test/tool', raw);
+
+      expect(result).not.toBeNull();
+      const text = result!.content[0]!.text;
+      expect(text).toContain('AI-generated summaries');
+      expect(text).toContain('Config section');
+      expect(text).toContain('Data section');
+    });
+
     it('falls back to simple index on LLM failure', async () => {
       const provider: LlmProvider = {
         name: 'test',
@@ -225,6 +244,56 @@ describe('ResponsePaginator', () => {
       const text = result!.content[0]!.text;
       expect(text).not.toContain('AI-generated summaries');
     });
+
+    it('passes modelOverride to provider.complete()', async () => {
+      const completeFn = vi.fn().mockResolvedValue({
+        content: JSON.stringify([{ page: 1, summary: 'test' }, { page: 2, summary: 'test2' }]),
+      });
+      const provider: LlmProvider = {
+        name: 'test',
+        isAvailable: () => true,
+        complete: completeFn,
+      };
+      const registry = {
+        getActive: () => provider,
+        register: vi.fn(),
+        setActive: vi.fn(),
+        listProviders: () => [{ name: 'test', available: true, active: true }],
+      } as unknown as ProviderRegistry;
+
+      const paginator = new ResponsePaginator(registry, { sizeThreshold: 100, pageSize: 60 }, 'gemini-2.5-pro');
+      const raw = makeLargeStringWithNewlines(150);
+      await paginator.paginate('test/tool', raw);
+
+      expect(completeFn).toHaveBeenCalledOnce();
+      const call = completeFn.mock.calls[0]![0]!;
+      expect(call.model).toBe('gemini-2.5-pro');
+    });
+
+    it('omits model when no modelOverride set', async () => {
+      const completeFn = vi.fn().mockResolvedValue({
+        content: JSON.stringify([{ page: 1, summary: 'test' }, { page: 2, summary: 'test2' }]),
+      });
+      const provider: LlmProvider = {
+        name: 'test',
+        isAvailable: () => true,
+        complete: completeFn,
+      };
+      const registry = {
+        getActive: () => provider,
+        register: vi.fn(),
+        setActive: vi.fn(),
+        listProviders: () => [{ name: 'test', available: true, active: true }],
+      } as unknown as ProviderRegistry;
+
+      const paginator = new ResponsePaginator(registry, { sizeThreshold: 100, pageSize: 60 });
+      const raw = makeLargeStringWithNewlines(150);
+      await paginator.paginate('test/tool', raw);
+
+      expect(completeFn).toHaveBeenCalledOnce();
+      const call = completeFn.mock.calls[0]![0]!;
+      expect(call.model).toBeUndefined();
+    });
   });
 
   // --- getPage ---

src/mcplocal/tests/project-mcp-endpoint.test.ts

@@ -6,8 +6,18 @@ import { registerProjectMcpEndpoint } from '../src/http/project-mcp-endpoint.js'
 // Mock discovery module — we don't want real HTTP calls
 vi.mock('../src/discovery.js', () => ({
   refreshProjectUpstreams: vi.fn(async () => ['mock-server']),
+  fetchProjectLlmConfig: vi.fn(async () => ({})),
 }));
 
+// Mock config module — don't read real config files
+vi.mock('../src/http/config.js', async () => {
+  const actual = await vi.importActual<typeof import('../src/http/config.js')>('../src/http/config.js');
+  return {
+    ...actual,
+    loadProjectLlmOverride: vi.fn(() => undefined),
+  };
+});
+
 import { refreshProjectUpstreams } from '../src/discovery.js';
 
 function mockMcpdClient() {