Skip to main content
Glama

start_session

Launch a Tauri app with configurable features, devtools, and timeout. Optionally wait until the application is ready.

Instructions

Start session (launch Tauri app)

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
wait_for_readyNoWait for ready
timeout_secsNoTimeout seconds
featuresNoCargo features to enable
devtoolsNoOpen devtools on launch

Implementation Reference

  • The handler function for start_session. It calls tauriManager.launch(args) to launch the Tauri app, then optionally sets a window title prefix from env var TAURI_MCP_WINDOW_PREFIX. Returns the launch result as JSON.
    start_session: async (args: { wait_for_ready?: boolean; timeout_secs?: number; features?: string[]; devtools?: boolean }) => {
      const result = await tauriManager.launch(args);
    
      // Apply window title prefix after successful launch if env var is set
      if (result.status === 'launched') {
        const prefix = process.env.TAURI_MCP_WINDOW_PREFIX;
        if (prefix) {
          try {
            await socketManager.setTitlePrefix(prefix);
          } catch {
            // Best-effort — app may not have finished initializing windows yet
          }
        }
      }
    
      return {
        content: [
          {
            type: 'text' as const,
            text: JSON.stringify(result, null, 2),
          },
        ],
      };
    },
  • The Zod schema definition for start_session, defining input parameters: wait_for_ready (optional boolean), timeout_secs (optional number), features (optional string array), and devtools (optional boolean).
    start_session: {
      name: 'start_session',
      description: 'Start session (launch Tauri app)',
      inputSchema: z.object({
        wait_for_ready: z.boolean().optional().describe('Wait for ready'),
        timeout_secs: z.number().optional().describe('Timeout seconds'),
        features: z.array(z.string()).optional().describe('Cargo features to enable'),
        devtools: z.boolean().optional().describe('Open devtools on launch'),
      }),
    },
  • Registration of start_session in the DEFAULT_ESSENTIAL_TOOLS array, which determines the set of tools available to clients unless overridden by the ESSENTIAL_TOOLS env var.
    const DEFAULT_ESSENTIAL_TOOLS = [
      'get_session_status',
      'start_session',
      'stop_session',
  • The launch() method on TauriManager called by the start_session handler. It handles port detection/allocation, checks for existing running app instances (idempotent), spawns 'pnpm tauri dev', optionally waits for the app to be ready, and returns a LaunchResult.
    async launch(options: LaunchOptions = {}): Promise<LaunchResult> {
      const waitForReady = options.wait_for_ready ?? true;
      const devtools = options.devtools ?? false;
    
      // Handle features as string or array (MCP may pass string)
      let features: string[] = [];
      if (options.features) {
        if (Array.isArray(options.features)) {
          features = options.features;
        } else if (typeof options.features === 'string') {
          features = (options.features as string).split(',').map(f => f.trim()).filter(Boolean);
        }
      }
    
      // Fallback to TAURI_MCP_FEATURES env if no features specified
      if (features.length === 0 && process.env.TAURI_MCP_FEATURES) {
        features = process.env.TAURI_MCP_FEATURES.split(',').map(f => f.trim()).filter(Boolean);
        console.error(`[tauri-mcp] Using default features from TAURI_MCP_FEATURES: ${features.join(',')}`);
      }
    
      if (!this.appConfig) {
        throw new Error('No Tauri app detected. Make sure src-tauri/Cargo.toml exists.');
      }
    
      // Idempotent: if already running (managed by this instance), return current status
      if (this.process) {
        const errors = this.parseBackendLogs(this.outputBuffer);
        const backendHealth = errors.some(e => e.level === 'error') ? 'error' as const : 'healthy' as const;
        return {
          status: 'already_running',
          message: 'App is already running',
          port: this.vitePort,
          portOverrideApplied: true, // Was applied on initial launch
          buildHealth: {
            frontend: 'unknown', // Will be determined by frontend logs
            backend: backendHealth,
          },
          errors: errors.filter(e => e.level === 'error'),
        };
      }
    
      // Check if another instance already has an app running (external process)
      // Auto-kill stale app instead of throwing — the socket is worktree-scoped,
      // so any app responding on it belongs to this same worktree's stale session
      const externalAppRunning = await this.checkExternalAppRunning();
      if (externalAppRunning) {
        console.error('[tauri-mcp] Stale external app detected on socket, force-stopping...');
        await this.forceStopExternalApp();
    
        // Verify the socket is gone after force-stop
        const stillAlive = await this.checkExternalAppRunning();
        if (stillAlive) {
          throw new Error(
            'Failed to stop the stale Tauri app instance on the socket. ' +
            'The process may require manual termination. ' +
            'Try: lsof -t ' + this.getSocketPath() + ' | xargs kill -9'
          );
        }
        console.error('[tauri-mcp] Stale app cleared, proceeding with launch');
      }
    
      // Determine timeout based on build cache existence
      // Fresh build: 300 seconds (5 minutes), Incremental build: 60 seconds
      const hasCachedBuild = this.hasBuildCache();
      const defaultTimeout = hasCachedBuild ? 60 : 300;
      const timeoutSecs = options.timeout_secs ?? defaultTimeout;
    
      console.error(`[tauri-mcp] Build cache ${hasCachedBuild ? 'found' : 'not found'}, using ${timeoutSecs}s timeout`);
    
      // Reset detected paths
      this.detectedPipePath = null;
      this.detectedUnixSocketPath = null;
    
      // Clean up stale socket file (Unix only)
      if (process.platform !== 'win32') {
        const socketPath = this.getSocketPath();
        if (fs.existsSync(socketPath)) {
          fs.unlinkSync(socketPath);
        }
      }
    
      // Dynamic port allocation with bundler detection
      const tauriConfig = this.readTauriConfig();
      let portOverrideApplied = false;
      const warnings: string[] = [];
      let configOverride: string | null = null;
    
      if (tauriConfig?.beforeDevCommand) {
        const bundlerType = this.detectBundlerType(tauriConfig.beforeDevCommand);
    
        if (bundlerType === 'vite') {
          // Vite supported - apply dynamic port
          this.vitePort = await this.findAvailablePort();
          const modifiedCommand = this.injectPortToCommand(tauriConfig.beforeDevCommand, this.vitePort);
          configOverride = JSON.stringify({
            build: {
              beforeDevCommand: modifiedCommand,
              devUrl: `http://localhost:${this.vitePort}`,
            },
          });
          portOverrideApplied = true;
          console.error(`[tauri-mcp] Vite detected. Using dynamic port ${this.vitePort}`);
        } else if (bundlerType === 'webpack') {
          // Webpack - try port override with warning
          this.vitePort = await this.findAvailablePort();
          const modifiedCommand = this.injectPortToCommand(tauriConfig.beforeDevCommand, this.vitePort);
          configOverride = JSON.stringify({
            build: {
              beforeDevCommand: modifiedCommand,
              devUrl: `http://localhost:${this.vitePort}`,
            },
          });
          portOverrideApplied = true;
          warnings.push(`Webpack detected. Port override may not work correctly. Using port ${this.vitePort}`);
          console.error(`[tauri-mcp] Webpack detected. Attempting dynamic port ${this.vitePort} (may not work)`);
        } else {
          // Unknown bundler - use default configuration
          this.vitePort = this.detectExistingPort() ?? 1420;
          warnings.push(
            `Unknown bundler in beforeDevCommand: "${tauriConfig.beforeDevCommand}". ` +
            `Dynamic port override not applied. Using default port ${this.vitePort}. ` +
            `If running multiple apps, port conflicts may occur.`
          );
          console.error(`[tauri-mcp] Unknown bundler. Using default port ${this.vitePort}`);
        }
      } else {
        // No beforeDevCommand - use default configuration
        this.vitePort = this.detectExistingPort() ?? 1420;
        warnings.push('No beforeDevCommand found in tauri.conf.json. Using default port configuration.');
        console.error(`[tauri-mcp] No beforeDevCommand. Using default port ${this.vitePort}`);
      }
    
      console.error(`[tauri-mcp] Launching app with Vite port ${this.vitePort}...`);
    
      // Build tauri dev command with optional features
      const tauriArgs = ['tauri', 'dev'];
      if (features.length > 0) {
        tauriArgs.push('--features', features.join(','));
      }
      // Add config override for dynamic port
      if (configOverride) {
        tauriArgs.push('--config', configOverride);
      }
      console.error(`[tauri-mcp] Command: pnpm ${tauriArgs.join(' ')}`);
      console.error(`[tauri-mcp] Socket will be at: ${this.getSocketPath()}`);
      this.process = spawn('pnpm', tauriArgs, {
        cwd: this.appConfig.appDir,
        stdio: ['ignore', 'pipe', 'pipe'],
        env: {
          ...process.env,
          // Use absolute appDir as project root for Rust plugin - this is where socket will be created
          TAURI_MCP_PROJECT_ROOT: path.resolve(this.appConfig.appDir),
          TAURI_MCP_DEVTOOLS: devtools ? '1' : '',
          // Pass window prefix for worktree identification
          ...(process.env.TAURI_MCP_WINDOW_PREFIX ? { TAURI_MCP_WINDOW_PREFIX: process.env.TAURI_MCP_WINDOW_PREFIX } : {}),
          // Keep VITE_PORT for backwards compatibility
          VITE_PORT: this.vitePort.toString(),
        },
        detached: process.platform !== 'win32',
        shell: process.platform === 'win32',
      });
    
    
    
      // Reset output buffer for this launch
      this.outputBuffer = [];
    
      this.process.stdout?.on('data', (data) => {
        const line = data.toString().trim();
        console.error(`[tauri stdout] ${line}`);
        this.outputBuffer.push(`[stdout] ${line}`);
        // Keep only last 100 lines
        if (this.outputBuffer.length > 100) this.outputBuffer.shift();
        // Parse Rust rebuild trigger
        this.parseRustRebuildTrigger(line);
      });
    
      this.process.stderr?.on('data', (data) => {
        const line = data.toString().trim();
        console.error(`[tauri stderr] ${line}`);
        this.outputBuffer.push(`[stderr] ${line}`);
        if (this.outputBuffer.length > 100) this.outputBuffer.shift();
        // Parse Rust rebuild trigger
        this.parseRustRebuildTrigger(line);
      });
    
      this.process.on('exit', (code) => {
        console.error(`[tauri-mcp] Process exited with code ${code}`);
        this.outputBuffer.push(`[exit] Process exited with code ${code}`);
        this.process = null;
    
        this.launchedAt = null;
        this.cleanupSocketFile();
      });
    
      if (waitForReady) {
        try {
          await this.waitForReady(timeoutSecs);
          const errors = this.parseBackendLogs(this.outputBuffer);
          const hasErrors = errors.some(e => e.level === 'error');
    
          this.launchedAt = Date.now();
          return {
            status: hasErrors ? 'build_error' : 'launched',
            message: hasErrors ? 'App started with build errors' : 'App is ready',
            port: this.vitePort,
            portOverrideApplied,
            warnings: warnings.length > 0 ? warnings : undefined,
            buildHealth: {
              frontend: 'unknown', // Will be determined by frontend logs
              backend: hasErrors ? 'error' : 'healthy',
            },
            errors: hasErrors ? errors.filter(e => e.level === 'error') : undefined,
          };
        } catch (e) {
          // Timeout or crash - still return useful info
          const errors = this.parseBackendLogs(this.outputBuffer);
          return {
            status: 'build_error',
            message: e instanceof Error ? e.message : 'Build failed',
            port: this.vitePort,
            portOverrideApplied,
            warnings: warnings.length > 0 ? warnings : undefined,
            buildHealth: {
              frontend: 'unknown',
              backend: 'error',
            },
            errors: errors.filter(e => e.level === 'error'),
          };
        }
      }
    
      this.launchedAt = Date.now();
      return {
        status: 'launched',
        message: 'App launched (not waiting for ready)',
        port: this.vitePort,
        portOverrideApplied,
        warnings: warnings.length > 0 ? warnings : undefined,
        buildHealth: {
          frontend: 'unknown',
          backend: 'unknown',
        },
      };
    }
  • Error message in socket.ts referencing start_session: when the socket file doesn't exist (ENOENT), the error advises using start_session first.
        client.on('error', (err) => {
          clearTimeout(timeout);
          if ((err as NodeJS.ErrnoException).code === 'ENOENT') {
            reject(new Error('App not running. Use start_session first.'));
          } else if ((err as NodeJS.ErrnoException).code === 'ECONNREFUSED') {
            reject(new Error('App is starting up. Please wait and try again.'));
          } else {
            reject(new Error(`Socket error: ${err.message}`));
          }
        });
    
        client.on('close', () => {
          if (settled) return;
          clearTimeout(timeout);
          reject(new Error(data
            ? 'Connection closed with incomplete response'
            : 'Connection closed without response'));
        });
      });
    }
Behavior2/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description must disclose behavioral traits. It only says 'launch Tauri app', missing critical details like whether the call blocks until the app is ready, if it returns a session ID, or what happens on failure.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Single sentence with no wasted words. The parenthetical adds clarity. Could be slightly more informative without losing conciseness, but it is not overly terse.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness2/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a tool that launches an application, the description lacks information about return values, error states, expected duration, and side effects. No output schema, so description should cover these aspects.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so parameters are documented. However, the tool description adds no meaning beyond the schema; for example, it doesn't explain how 'wait_for_ready' interacts with 'timeout_secs'. Baseline of 3 is appropriate.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the verb ('Start') and the resource ('session'), and includes a parenthetical explanation ('launch Tauri app') that distinguishes it from siblings like 'stop_session' and 'get_session_status'. However, 'session' could be more specific about what kind of session.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

No guidance is provided on when to use this tool versus alternatives. For instance, it doesn't explain whether it should be called before other tools, if there are prerequisites, or when 'stop_session' is needed.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/DaveDev42/tauri-plugin-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server