FastMCP

The fast, Pythonic way to build MCP servers and clients.

Docs

• Architecture: How FastMCP apps work under the hood — from Python to pixels.
• Development: Preview and test your app tools locally without a full MCP host.
• Examples: Example apps you can run right now.
• FastMCPApp: Wire an interactive UI to backend tools with managed visibility and composition safety.
• Generative UI: Let the LLM build custom Prefab UIs on the fly.
• Custom HTML Apps: Build apps with your own HTML, CSS, and JavaScript using the MCP Apps extension directly.
• Apps: Give your tools interactive UIs rendered directly in the conversation.
• Interactive Tools: Turn your tools into interactive UIs with charts, tables, and dashboards.
• Approval: Human-in-the-loop approval gates for agent actions
• Choice: Present clickable options instead of free-text responses
• File Upload: Drag-and-drop file upload for any MCP server
• Form Input: Collect structured data from users via Pydantic models
• Quickstart: Build your first FastMCP app in under a minute.
• Changelog
• Auth Utilities: Create and validate CIMD documents for OAuth
• Client Commands: List tools, call them, and discover configured servers
• Generate CLI: Scaffold a standalone typed CLI from any MCP server
• Inspecting Servers: View a server's components and metadata
• Install MCP Servers: Install MCP servers into Claude, Cursor, Gemini, and other clients
• CLI: The fastmcp command-line interface
• Running Servers: Start, develop, and configure servers from the command line
• Bearer Token Authentication: Authenticate your FastMCP client with a Bearer token.
• CIMD Authentication: Use Client ID Metadata Documents for verifiable, domain-based client identity.
• OAuth Authentication: Authenticate your FastMCP client via OAuth 2.1.
• The FastMCP Client: Programmatic client for interacting with MCP servers through a well-typed, Pythonic interface.
• Client-Only Package: Use FastMCP's client without installing the full server framework.
• User Elicitation: Handle server requests for structured user input.
• fastmcp-remote: Bridge remote MCP servers into stdio-only MCP hosts with uvx fastmcp-remote.
• Server Logging: Receive and handle log messages from MCP servers.
• Notifications: Handle server-sent notifications for list changes and other events.
• Progress Monitoring: Handle progress notifications from long-running server operations.
• Getting Prompts: Retrieve rendered message templates with automatic argument serialization.
• Reading Resources: Access static and templated data sources from MCP servers.
• Client Roots: Provide local context and resource boundaries to MCP servers.
• LLM Sampling: Handle server-initiated LLM completion requests.
• Background Tasks: Execute operations asynchronously and track their progress.
• Calling Tools: Execute server-side tools and handle structured results.
• Client Transports: Configure how clients connect to and communicate with MCP servers.
• HTTP Deployment: Deploy your FastMCP server over HTTP for remote access
• Prefect Horizon: The MCP platform from the FastMCP team
• Running Your Server: Learn how to run your FastMCP server locally for development and testing
• Sandboxed Agents: Expose MCP tools to isolated agents without giving the sandbox long-lived credentials.
• Project Configuration: Use fastmcp.json for portable, declarative project configuration
• Contributing: Development workflow for FastMCP contributors
• Releases: FastMCP versioning and release process
• Tests: Testing patterns and requirements for FastMCP
• Installation: Install FastMCP and verify your setup
• Quickstart
• Upgrading from FastMCP 2: Migration instructions for upgrading between FastMCP versions
• Upgrading from the MCP Low-Level SDK: Upgrade your MCP server from the low-level Python SDK's Server class to FastMCP
• Upgrading from the MCP SDK: Upgrade from FastMCP in the MCP Python SDK to the standalone FastMCP framework
• Welcome to FastMCP: The fast, Pythonic way to build MCP servers, clients, and applications.
• Anthropic API 🤝 FastMCP: Connect FastMCP servers to the Anthropic API
• Auth0 OAuth 🤝 FastMCP: Secure your FastMCP server with Auth0 OAuth
• AuthKit 🤝 FastMCP: Secure your FastMCP server with AuthKit by WorkOS
• AWS Cognito OAuth 🤝 FastMCP: Secure your FastMCP server with AWS Cognito user pools
• Azure (Microsoft Entra ID) OAuth 🤝 FastMCP: Secure your FastMCP server with Azure/Microsoft Entra OAuth
• ChatGPT 🤝 FastMCP: Connect FastMCP servers to ChatGPT in Chat and Deep Research modes
• Claude Code 🤝 FastMCP: Install and use FastMCP servers in Claude Code
• Claude Desktop 🤝 FastMCP: Connect FastMCP servers to Claude Desktop
• Cursor 🤝 FastMCP: Install and use FastMCP servers in Cursor
• Descope 🤝 FastMCP: Secure your FastMCP server with Descope
• Discord OAuth 🤝 FastMCP: Secure your FastMCP server with Discord OAuth
• Eunomia Authorization 🤝 FastMCP: Add policy-based authorization to your FastMCP servers with Eunomia
• FastAPI 🤝 FastMCP: Integrate FastMCP with FastAPI applications
• Gemini SDK 🤝 FastMCP: Connect FastMCP servers to the Google Gemini SDK
• Gemini CLI 🤝 FastMCP: Install and use FastMCP servers in Gemini CLI
• GitHub OAuth 🤝 FastMCP: Secure your FastMCP server with GitHub OAuth
• Google OAuth 🤝 FastMCP: Secure your FastMCP server with Google OAuth
• Goose 🤝 FastMCP: Install and use FastMCP servers in Goose
• Keycloak OAuth 🤝 FastMCP: Secure your FastMCP server with Keycloak OAuth
• MCP JSON Configuration 🤝 FastMCP: Generate standard MCP configuration files for any compatible client
• OCI IAM OAuth 🤝 FastMCP: Secure your FastMCP server with OCI IAM OAuth
• OpenAI API 🤝 FastMCP: Connect FastMCP servers to the OpenAI API
• OpenAPI 🤝 FastMCP: Generate MCP servers from any OpenAPI specification
• Permit.io Authorization 🤝 FastMCP: Add fine-grained authorization to your FastMCP servers with Permit.io
• PropelAuth 🤝 FastMCP: Secure your FastMCP server with PropelAuth
• Pydantic AI 🤝 FastMCP: Connect FastMCP servers to Pydantic AI agents using the FastMCPToolset
• Scalekit 🤝 FastMCP: Secure your FastMCP server with Scalekit
• Supabase 🤝 FastMCP: Secure your FastMCP server with Supabase Auth
• WorkOS 🤝 FastMCP: Authenticate FastMCP servers with WorkOS Connect
• FAQ: Answers to common questions about installing and using FastMCP
• Settings: Configure FastMCP behavior through environment variables or a .env file.
• Contrib Modules: Community-contributed modules extending FastMCP
• init
• app
• approval
• choice
• config
• file_upload
• form
• generative
• cli
• client
• decorators
• dependencies
• exceptions
• init
• init
• handlers
• init
• code_mode
• mcp_config
• prompts
• resources
• server
• settings
• telemetry
• tools
• types
• init
• async_utils
• auth
• authorization
• cli
• components
• docstring_parsing
• exceptions
• http
• inspect
• json_schema
• json_schema_type
• lifespan
• logging
• init
• init
• init
• base
• uv
• mcp_server_config
• init
• base
• filesystem
• mime
• openapi
• pagination
• skills
• tasks
• tests
• timeout
• token_cache
• types
• ui
• version_check
• versions
• Authentication: Secure your FastMCP server with flexible authentication patterns, from simple API keys to full OAuth 2.1 integration with external identity providers.
• Full OAuth Server: Build a self-contained authentication system where your FastMCP server manages users, issues tokens, and validates them.
• Multiple Auth Sources: Accept tokens from multiple authentication sources with a single server.
• OAuth Proxy: Bridge traditional OAuth providers to work seamlessly with MCP's authentication flow.
• OIDC Proxy: Bridge OIDC providers to work seamlessly with MCP's authentication flow.
• Remote OAuth: Integrate your FastMCP server with external identity providers like Descope, WorkOS, Auth0, and corporate SSO systems.
• Token Verification: Protect your server by validating bearer tokens issued by external systems.
• Authorization: Control access to components using callable-based authorization checks that filter visibility and enforce permissions.
• Composing Servers: Combine multiple servers into one
• MCP Context: Access MCP capabilities like logging, progress, and resources within your MCP objects.
• Dependency Injection: Inject runtime values like HTTP requests, access tokens, and custom dependencies into your MCP components.
• User Elicitation: Request structured input from users during tool execution through the MCP context.
• Icons: Add visual icons to your servers, tools, resources, and prompts
• Lifespans: Server-level setup and teardown with composable lifespans
• Client Logging: Send log messages back to MCP clients through the context.
• Middleware: Add cross-cutting functionality to your MCP server with middleware that intercepts and modifies requests and responses.
• Pagination: Control how servers return large lists of components to clients.
• Progress Reporting: Update clients on the progress of long-running operations through the MCP context.
• Prompts: Create reusable, parameterized prompt templates for MCP clients.
• Custom Providers: Build providers that source components from any data source
• Filesystem Provider: Automatic component discovery from Python files
• Local Provider: The default provider for decorator-registered components
• Providers: How FastMCP sources tools, resources, and prompts
• MCP Proxy Provider: Source components from other MCP servers
• Skills Provider: Expose agent skills as MCP resources
• Resources & Templates: Expose data sources and dynamic content generators to your MCP client.
• Sampling: Request LLM text generation from the client or a configured provider through the MCP context.
• The FastMCP Server: The core FastMCP server class for building MCP applications
• Storage Backends: Configure persistent and distributed storage for caching and OAuth state management
• Background Tasks: Run long-running operations asynchronously with progress tracking
• OpenTelemetry: Native OpenTelemetry instrumentation for distributed tracing.
• Testing your FastMCP Server: How to test your FastMCP server.
• Tool Fingerprinting: Build stable fingerprints for tool identity and schema change detection
• Tools: Expose functions as executable capabilities for your MCP client.
• Code Mode: Let LLMs write Python to orchestrate tools in a sandbox
• Namespace Transform: Prefix component names to prevent conflicts
• Prompts as Tools: Expose prompts to tool-only clients
• Resources as Tools: Expose resources to tool-only clients
• Tool Search: Replace large tool catalogs with on-demand search
• Tool Transformation: Modify tool schemas - rename, reshape arguments, and customize behavior
• Transforms Overview: Modify components as they flow through your server
• Versioning: Serve multiple API versions from a single codebase
• Component Visibility: Control which components are available to clients
• FastMCP Updates