Smithery

Smithery MCP Connector for Claude

A+

AI MCP registry: discover, search, and connect MCP servers to your agents via Smithery.

11 tools Official Updated Jun 28, 2026 Official Vinkius Partner

What you can do

Connect AI agents to the Smithery Registry for comprehensive MCP server discovery and management:

  • Search MCP servers — find servers by name, description, or tags with semantic search
  • Get server details — review metadata, verification status, and user counts
  • Discover tools — list all tools (functions) exposed by any registered MCP server
  • Discover resources — list all data resources available from MCP servers
  • Discover prompts — list all prompt templates exposed by MCP servers
  • Create connections — connect to MCP servers via Smithery Connect with automatic OAuth handling
  • Manage connections — list, inspect, and remove MCP server connections
  • Generate service tokens — create scoped, time-limited tokens for frontend/agent access
  • View analytics — monitor server usage, adoption trends, and performance metrics

How it works

  1. Get your Smithery API key from smithery.ai/account/api-keys
  2. Ask your AI agent to search the registry, discover server capabilities, or create connections
  3. Natural language commands replace manual Smithery dashboard navigation
  4. Zero OAuth configuration — Smithery Connect handles tokens and sessions automatically

Who is this for?

Essential for AI application developers, agent architects, platform engineers, MCP server publishers, and enterprise AI teams building with the Model Context Protocol. Let AI agents discover relevant MCP servers, review their capabilities, create managed connections, and monitor usage analytics. Perfect for teams integrating multiple MCP servers who want to streamline server discovery, eliminate manual OAuth configuration, and enable AI-driven MCP ecosystem management.

registryapi-managementdiscoveryoauthanalytics

11 tools expose this connector's capabilities to your AI agent.

create_connection

Smithery handles OAuth, tokens, and sessions automatically. Requires the server namespace and connection configuration (mcpUrl, optional headers, metadata). Returns the connection ID, status, and server info. Use this to integrate MCP servers into your applications without managing authentication complexity. Create a new connection to an MCP server via Smithery Connect

create_service_token

The token has limited permissions defined by the policy (namespaces, resources, operations, metadata, TTL). Returns the token string. Use this to provide secure, time-limited access to MCP servers without exposing your main API key. Generate a scoped service token for frontend/agent access to MCP servers

delete_connection

This action cannot be undone. Requires namespace and connection ID. Use this to clean up unused connections or revoke access. Remove an MCP server connection

get_connection

Requires namespace and connection ID. Use this to review connection details or troubleshoot connectivity issues. Get detailed information about a specific MCP connection

get_server_analytics

Requires the server qualified name. Use this to monitor server adoption, identify usage trends, or troubleshoot performance issues. Get usage analytics for a specific MCP server

get_server_details

Requires the qualified name (e.g., "smithery/hello-world" or "github/github") from search_servers results. Use this to review server capabilities before connecting. Get detailed information about a specific MCP server from the Smithery registry

get_server_prompts

Returns prompt names, descriptions, and argument definitions. Requires the server qualified name. Use this to discover reusable prompt workflows available from the server. List all prompt templates exposed by a specific MCP server

get_server_resources

Returns resource URIs, names, descriptions, and MIME types. Requires the server qualified name. Use this to understand what data the server provides read access to. List all resources exposed by a specific MCP server

get_server_tools

Returns tool names, descriptions, input schemas, and annotations. Requires the server qualified name. Use this to understand what actions the server can perform before connecting it to your agents. List all tools exposed by a specific MCP server

list_connections

Returns connection IDs, names, statuses, creation dates, and metadata. Use this to audit which connections are active, review connection configurations, or identify unused connections. List all connections for a specific MCP server namespace

search_servers

Returns matching servers with qualified names, descriptions, verification status, user counts, and deployment info. Use optional filters to narrow by namespace, verified status, or deployment state. Results include pagination metadata. Use this as the first step to discover available MCP servers before connecting or installing them. Search the Smithery registry for MCP servers by name, description, or tags

See how to talk to your AI agent using Smithery.

Search for verified GitHub-related MCP servers

I'll search the Smithery registry for verified GitHub MCP servers.

Show me all tools exposed by the Stripe MCP server

I'll retrieve the complete tool catalog from the Stripe MCP server.

Create a connection to the Slack MCP server for my workspace

I'll create a managed connection to the Slack MCP server through Smithery Connect.

Yes, you need a free Smithery account to generate an API key. Sign up at smithery.ai, then go to Account Settings > API Keys to create your key. The key gives access to the registry search, server details, and connection management features.

Related Connectors