tauri-docs

Mastra MCP server providing access to Tauri documentation from tauri.app.

📦 Versioning & Releases

This project uses semantic-release for automated versioning and changelog generation based on conventional commits.

Commit Message Format

Use conventional commit format for automatic versioning:

• feat: - Minor version bump (new features)

• fix: - Patch version bump (bug fixes)

• BREAKING CHANGE: - Major version bump (breaking changes)

• docs:, chore:, refactor:, test:, ci: - No version bump

Release Triggers

Releases are automatically triggered on pushes to main branch when commit messages contain version keywords:

• [patch] - Forces patch release

• [minor] - Forces minor release

• [major] - Forces major release

Alternatively, trigger manually via GitHub Actions → "Release Changelog" → "Run workflow".

What Happens on Release

1. Analyzes commits since last release

2. Determines version bump based on conventional commits

3. Generates/updates CHANGELOG.md

4. Creates git tag (e.g., v1.0.0)

5. Commits changes back to repository

Viewing Changes

See CHANGELOG.md for detailed release notes and history.

Production Deployments

Choose the base host that fits your workflow — both expose the same toolset, but their runtime characteristics differ:

• Append /api/mcp/tauri-docs/sse for the SSE transport (best for editors that keep long-lived connections).

• Append /api/mcp/tauri-docs/mcp for the HTTP transport (handy for CLIs and quick one-off calls).

• Mastra Cloud is the recommended primary deployment - it offers zero cold start and maximum responsiveness.

Endpoint reference & alternates

• Mastra Cloud SSE: https://tauri-docs.mastra.cloud/api/mcp/tauri-docs/sse

• Mastra Cloud HTTP: https://tauri-docs.mastra.cloud/api/mcp/tauri-docs/mcp

This repository contains a Mastra-based MCP server that provides access to Tauri documentation from tauri.app. Use it in your AI-powered code editor to get instant access to the latest Tauri documentation directly from the official Tauri documentation site.

🎉 What's New

• ✅ MCP server deployed on Mastra Cloud

• ✅ Four main MCP tools for documentation discovery, page retrieval, and search

• ✅ Advanced LRU caching with automatic eviction and size limits

• ✅ Request metrics and health monitoring

• ✅ TypeScript type safety with Zod schemas

• ✅ Resources API for static documentation metadata

• ✅ Guided prompts for common Tauri workflows

• ✅ Support for all major AI code editors (Cursor, Windsurf, VS Code, Zed, Claude Code, Codex)

• ✅ HTTP and SSE transport protocols

• ✅ Real-time web scraping from tauri.app

Mastra Cloud is the recommended deployment for reliability and responsiveness.

Windsurf

1. Edit ~/.codeium/windsurf/mcp_config.json.

2. Mastra Cloud is recommended for zero cold start and maximum responsiveness. Add the SSE transport as shown:

3. Save, restart Windsurf, then open mcp.json in Agent mode and click "start".

Use the HTTP variant if you need it:

Zed

1. Open Zed settings (Cmd/Ctrl + ,).

2. Edit ~/.config/zed/settings.json and add an entry under context_servers:

3. Save, restart Zed, and confirm the server shows a green indicator in the Agent panel. Zed also offers a UI flow via Settings → Agent to paste either endpoint without editing JSON.

Cursor

1. Open Cursor Settings (Cmd/Ctrl + ,).

2. Navigate to "MCP" / "Model Context Protocol" and add a new server configuration.

3. Mastra Cloud is recommended for zero cold start and maximum responsiveness. Append the SSE or HTTP path as shown in the examples below.

Mastra Cloud — SSE example:

Mastra Cloud — HTTP example:

VS Code

VS Code users can open the Command Palette (Cmd/Ctrl+Shift+P) and run MCP: Add server to paste either URL.

The same base URLs work across CLIs. Mastra Cloud is the recommended primary deployment for the fastest responses with zero cold start.

Claude Code CLI (Anthropic)

• Global settings (~/.claude/settings.json):

• Project-scoped override (.mcp.json):

Enable project servers with:

• Command palette alternative:

• Use /permissions inside Claude Code to grant tool access if prompted.

OpenAI Codex CLI

Register the Mastra Cloud endpoint for codex or use your own privately hosted MCP endpoint.

Gemini CLI (Google)

1. Create or edit ~/.gemini/settings.json:

2. Add a configuration. Mastra Cloud example:

3. Prefer the npx mcp-remote command variant if your CLI version expects a command:

4. Mastra Cloud is recommended for zero cold start and maximum responsiveness. Restart the CLI to apply changes.

Verification & Quick Tests

• claude mcp list

• codex mcp list

• npx mcp-remote https://tauri-docs.mastra.cloud/api/mcp/tauri-docs/mcp

• curl -I https://tauri-docs.mastra.cloud/api/mcp/tauri-docs/mcp

• curl -N https://tauri-docs.mastra.cloud/api/mcp/tauri-docs/sse

Claude Code may prompt for tool permissions — use /permissions or set allowedTools in ~/.claude.json. Editors that maintain long-lived connections should use the SSE URL; quick scripts can stick with HTTP.

Available Tools

Once installed, your AI assistant will have access to these tools (IDs exactly as exposed by the MCP server):

Core Tools

1. list_sections — Parse https://tauri.app/llms.txt to list doc sections

2. get_page — Fetch a Tauri doc page and return cleaned HTML content

3. search — Keyword search across the llms.txt index

4. get_plugin — Fetch plugin doc pages by name

Resources (NEW)

Static, auto-updating resources available via the MCP Resources API:

1. tauri://docs/structure — Complete documentation structure from llms.txt

2. tauri://platforms — Supported platforms (Windows, macOS, Linux, iOS, Android)

3. tauri://metrics — Real-time server metrics (requests, cache, health)

Prompts (NEW)

Guided workflows for common tasks:

1. getting-started — Step-by-step guide to create your first Tauri app

2. troubleshooting — Common issues and debugging workflows

3. plugin-setup — Guide to installing and configuring plugins

4. migration-v1-to-v2 — Guide for migrating from Tauri v1 to v2

Tool response formats (quick reference)

• list_sections: List of documentation sections from llms.txt with total count

• get_page: Cleaned HTML documentation for a specific page

• search: List of matching sections with relevance scores and total count

Contents

• src/ - Mastra bootstrap, MCP servers, tools, and agents

• src/mastra/tools/ - Tools for accessing Tauri documentation

• src/mastra/lib/ - Caching, parsing, metrics, and utility functions

  • types.ts - TypeScript types and Zod schemas

  • cache-manager.ts - LRU cache with automatic eviction

  • metrics.ts - Request tracking and health monitoring

  • llms-txt.ts - Documentation index parsing

  • html.ts - HTML fetching and cleaning

• scripts/ - Version management and automation scripts (if any)

• "Show me the Tauri plugin documentation"

• "Get the overview of Tauri APIs"

• "List all sections in Tauri docs"

• "Search for Tauri configuration options"

• "What are the methods available in Tauri?"

• "Find plugins related to web frameworks"

• "Get documentation for the Tauri window API"

• "Search for docs with 'security' in the name"

• "Show me the Tauri CLI documentation"

Local Development

Want to run the MCP server locally or contribute to the project?

Contents

• src/ - Mastra bootstrap, MCP servers, tools, and agents

• src/mastra/tools/ - Tools for accessing Tauri documentation

• src/mastra/lib/ - Caching, parsing, and utility functions

• scripts/ - Version management and automation scripts (if any)

Quick start (development smoke-test)

1. Install dependencies (using your preferred package manager).

2. Run the development smoke-test (recommended):

Useful scripts

• npm run dev – Start Mastra in development mode (recommended smoke-test).

• npm run build – Build the Mastra project for production.

• npm run start – Start the built Mastra server.

• npm run check-versions – Check if package.json and mcp-server.ts versions match (fails if mismatched). (If applicable)

• npm run sync-versions-auto – Check versions and auto-sync if mismatched (package.json is source of truth). (If applicable)

• npm run sync-versions – Sync versions from latest git tag to both files. (If applicable)

MCP Architecture

This project exposes a production-ready MCP Server that makes Tauri documentation available to AI code editors.

What this means:

• MCP Server (src/mastra/index.ts) - Exposes four Tauri documentation tools to external MCP clients (Cursor, Windsurf, VS Code, etc.)

• No MCP Client needed - This project only provides tools, it doesn't consume tools from other servers

The server is deployed at https://tauri-docs.mastra.cloud and exposes tools via HTTP and SSE transports.

Project Architecture

Key Features

• Real-time Documentation: Always fetches latest content from tauri.app

• Comprehensive Coverage: Access to guides, APIs, plugins, and references

• Advanced Caching: LRU cache with automatic eviction and size limits (10MB index, 50MB pages)

• Metrics & Monitoring: Request tracking, cache statistics, and health checks

• Type Safety: Full TypeScript support with Zod validation

• Resources API: Static documentation metadata (structure, platforms, plugins, metrics)

• Guided Prompts: Step-by-step workflows for common tasks

• Search Functionality: Find documentation by keywords with relevance scoring

• Clean HTML Output: Returns parsed documentation with navigation/scripts removed

Conventions & notes

• Tools use zod for input validation and follow Mastra patterns with createTool

• Web scraping uses cheerio and turndown for HTML to markdown conversion

• Intelligent caching reduces API calls while ensuring freshness

Development tips

• Node >= 22.13.0 required (see package.json engines)

• Run npm run dev for smoke tests after changes

• Clear cache during development if fresh data needed

License

This project is licensed under the MIT License - see the LICENSE file for details.

Contributing

We welcome contributions! Please read our Contributing Guidelines and Code of Conduct before getting started.

Contact

• Issues & Support: GitHub Issues

• Maintainer: Michael Obele

For more details:

• Web scraping services: See src/mastra/lib/ for documentation fetching and parsing implementation