I’m now building Executor full time as a startup! The state of tool calling is a mess: - Everyone is using different agents - You have no idea what actions are going to run - Destructive actions are treated the same as normal Executor lets you set up everything once & use it with any agent. You’re able to control what tools can be called, require approval on destructive actions, and invite your team to all use them. Under the hood, it’s built on codemode to let you add thousands of tools with 0 context bloat. I’ve been using it for a month so far and it has completely changed how I interact with agents. There’s so much to be built here, immediately next up on the roadmap is workspaces, skills, workflows, and generative ui. Everything is being built open source over at https://github.com/RhysSullivan/executor…, there’s a lot more coming on an SDK to embed it into your own agents, a self hostable cloud app, and plugins There’s an early beta version over at http://executor.sh. It’s very early on - if you hit any bugs let me know and I’ll get them fixed up, excited to share more as I build it!

RhysSullivan/executor

Source: https://github.com/RhysSullivan/executor

executor

https://github.com/user-attachments/assets/11225f83-e848-42ba-99b2-a993bcc88dad

The integration layer for AI agents. One catalog for every tool, shared across every agent you use.

Ask DeepWiki

Quick start

npm install -g executor
executor web

This starts a local runtime with a web UI at http://127.0.0.1:4788. From there, add your first source and start using tools.

Use as an MCP server

Point any MCP-compatible agent (Cursor, Claude Code, OpenCode, etc.) at Executor to share your tool catalog, auth, and policies across all of them.

executor mcp

Example mcp.json for Claude Code / Cursor:

{
  "mcpServers": {
    "executor": {
      "command": "executor",
      "args": ["mcp"]
    }
  }
}

Add a source

If you can represent it with a JSON schema, it can be an integration. Executor has first-party support for OpenAPI, GraphQL, MCP, and Google Discovery — but the plugin system is open to any source type.

Via the web UI

Open http://127.0.0.1:4788, go to Add Source, paste a URL, and Executor will detect the type, index the tools, and handle auth.

Via the CLI

executor call openapi addSource '{
  "spec": "https://petstore3.swagger.io/api/v3/openapi.json",
  "namespace": "petstore",
  "baseUrl": "https://petstore3.swagger.io/api/v3"
}'

Use baseUrl when the OpenAPI document has relative servers entries (for example "/api/v3").

Use tools

Agents discover and call tools through a typed TypeScript runtime:

// discover by intent
const matches = await tools.discover({ query: "github issues", limit: 5 });

// inspect the schema
const detail = await tools.describe.tool({
  path: matches.bestPath,
  includeSchemas: true,
});

// call with type safety
const issues = await tools.github.issues.list({
  owner: "vercel",
  repo: "next.js",
});

Use tools via the CLI:

executor tools search "send email"
executor call --help
executor call github --help
executor call github issues --help
executor call cloudflare --help --match dns --limit 20
executor call github issues create '{"owner":"octocat","repo":"Hello-World","title":"Hi"}'
executor call gmail send '{"to":"[email protected]","subject":"Hi"}'

executor call, executor resume, and executor tools ... commands auto-start a local daemon if needed. If the default port is busy, the CLI will pick an available local port and track it automatically.

If an execution pauses for auth or approval, resume it:

executor resume --execution-id exec_123

CLI reference

executor web                        # start runtime + web UI
executor daemon run                 # start persistent local daemon in background
executor daemon status              # show daemon status
executor daemon stop                # stop daemon
executor daemon restart             # restart daemon
executor mcp                        # start MCP endpoint
executor call <path...> '{"k":"v"}' # invoke a tool by path segments
executor call <path...> --help      # browse namespaces/resources/methods
executor call <path...> --help --match "<text>" --limit <n> # narrow huge namespaces
executor resume --execution-id <id> # resume paused execution
executor tools search "<query>"     # search tools by intent
executor tools sources              # list configured sources + tool counts
executor tools describe <path>      # show tool TypeScript/JSON schema

Developing locally

bun install
bun dev

The dev server starts at http://127.0.0.1:4788.

Community

Join the Discord: https://discord.gg/eF29HBHwM6

Learn more

Visit executor.sh to learn more.

Attribution

• Thank you to Crystian for providing the npm package name executor.

References

As part of my coding process, I give my agent access to references to other codebases to understand patterns and how other people have implemented systems.

A non exhaustive list of references are:

• Better Auth - Storage adapter reference
• Effect - General code patterns
• OpenCode - Plugin system reference
• OpenClaw - Plugin system reference
• Emdash - Plugin system reference
• Pi - Plugin system reference

It’s encouraged also that you can use this codebase as a reference to understand how it’s implemented