MCP adapters

Corsair provides adapters for common AI SDKs and agent runtimes. Each adapter exposes Corsair’s tools in the format that framework expects, so no manual schema wiring is required.

Available adapters

Adapter	Use case
Anthropic SDK	Native tool use with Claude models
Claude Agent SDK	In-process MCP with the Claude Agent SDK
OpenAI Agents	OpenAI Agents SDK tool integration
OpenAI	OpenAI function calling
Vercel AI SDK	Tools for `useChat` and `streamText`
Mastra	Mastra agent tool integration

Coding agents

For coding agents that use the MCP stdio protocol, see the Coding Agents section.

Tools

Every adapter exposes the same three tools automatically:

Tool	What it does
`list_operations`	Discover every available API endpoint
`get_schema`	Inspect parameters for a specific endpoint
`run_script`	Execute a JS snippet with `corsair` in scope

Your agent calls list_operations to discover what’s available, get_schema to inspect it, then run_script to execute. No code changes needed as you add plugins.

How the agent uses Corsair

Once connected, your agent follows this pattern automatically:

list_operations        → discover available endpoints (github.repositories.list, slack.messages.post, ...)
get_schema             → inspect parameters for a specific endpoint
run_script             → execute: const repos = await corsair.github.api.repositories.list({ type: 'owner' })

No hard-coding required. As you add plugins, the agent discovers the new endpoints automatically.

Get Started

Frameworks

Adapters

Workflows (beta)

Reference

Available adapters

Coding agents

Tools

How the agent uses Corsair

​Available adapters

​Coding agents

​Tools

​How the agent uses Corsair

Available adapters

Coding agents

Tools

How the agent uses Corsair