Overview
The Layers MCP server gives AI agents native Layers tools for the API integration and keyless onboarding, with no HTTP code to write.
The Layers MCP server exposes the API integration as Model Context Protocol tools. An MCP client (Claude, ChatGPT via the OpenAI Agents SDK, Cursor, or any other) spawns it, and the agent can then create projects, generate content, schedule posts, and read metrics by calling tools directly — without you writing a single fetch.
It is published to npm as @layers/mcp-server and runs locally over stdio — you spawn it; there is no service for you to operate. One hosted MCP endpoint also exists: the keyless agent onboarding server, which a trial session connects to over HTTP before anyone has an account or an API key.
The API-backed surface is a thin wrapper. Its 52 tools each call one
documented /v1 endpoint. Five keyless tools sit outside that one-to-one
surface: four drive onboarding, and ask_elle routes the guided conversation
rather than calling /v1.
Who it's for
- Agent builders wiring Layers into an AI assistant or autonomous workflow, who want tool-calling instead of an HTTP client.
- Anyone prototyping against the API integration who'd rather say "list my Layers projects" in an MCP client than hand-roll requests.
If you're integrating Layers into a conventional backend, you don't need the MCP server — call the API integration directly.
How it relates to the API integration
With an lp_... key, the MCP server is a client of the API integration. It talks to the same https://api.layers.com/v1 host and inherits the same auth, scopes, idempotency, rate limits, and sandbox behavior.
Without a key, the package starts in keyless onboarding mode. By default, it registers five onboarding and guidance tools alongside the 52 workspace tools. The workspace tools refuse requests until the trial is claimed, then use the claimed workspace's credential internally.
For the 52 API-backed tools, everything you already know about the API carries over:
- Auth & scopes — the key you pass the server determines what it can do. See Authentication.
- Sandbox — a
lp_test_...key returns fixture-backed results, so an agent can be developed safely against it. See Sandbox. - Async jobs, pagination, idempotency, errors — identical semantics, surfaced through MCP conventions. See How it works.
Next steps
Quickstart
Connect the server to Claude or the OpenAI Agents SDK in under a minute.
Agent onboarding
The keyless path: proof of work, one POST, and a hosted MCP endpoint — no account, no API key.
Authentication & configuration
Flags and env vars: API key, base URL, read-only mode, child-org acting.
How it works
Transport, server instructions, tool annotations, async jobs, pagination, idempotency, errors.
Tool reference
The 52 API-backed tools, plus the five tools available in keyless mode.
Package
- npm:
@layers/mcp-server - Source: github.com/layers/mcp-server
- License: Apache-2.0