openclaw-mcp

openclaw-mcp is an MCP (Model Context Protocol) server/bridge that connects Claude (MCP client) to a self-hosted OpenClaw gateway/assistant. It forwards chat/status/task operations and supports sync and async tools, including multi-instance routing and OAuth2 authentication.

Evaluated Mar 30, 2026 (21d ago)

Homepage ↗ Repo ↗ Ai Ml mcp mcp-server openclaw claude oauth2 bridge self-hosted typescript security docker

⚙ Agent Friendliness

/ 100

Can an agent use this?

🔒 Security

/ 100

Is it safe for agents?

⚡ Reliability

/ 100

Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality

Documentation

Error Messages

Auth Simplicity

Rate Limits

🔒 Security

TLS Enforcement

Auth Strength

Scope Granularity

Dep. Hygiene

Secret Handling

Uses OAuth2 authentication for the MCP server (AUTH_ENABLED=true) and provides guidance to restrict CORS origins. README includes a threat model and logging docs (referenced but not included here). Scope granularity and specific authorization model are not detailed; rate limiting details are not provided. Dependency hygiene cannot be fully assessed from the provided manifest/README.

⚡ Reliability

Uptime/SLA

Version Stability

Breaking Changes

Error Recovery

Best When

You want a straightforward MCP integration for Claude with OAuth2-protected access to a self-hosted OpenClaw backend, including optional multi-instance routing.

Avoid When

You cannot set correct OAuth issuer/public URLs behind a proxy or you plan to run the service with authentication disabled in production.

Use Cases

• Connect Claude.ai or Claude Desktop to a self-hosted OpenClaw assistant via MCP
• Send chat messages to OpenClaw and retrieve responses
• Run long-running OpenClaw tasks asynchronously (queue, poll status, cancel)
• Monitor OpenClaw gateway health and list configured OpenClaw instances
• Bridge to multiple OpenClaw gateway instances (prod/staging/dev) from a single MCP endpoint

Not For

• Exposing OpenClaw capabilities publicly without proper authentication/authorization controls
• Use cases requiring a documented REST/GraphQL/SDK interface beyond MCP tooling
• Environments where you cannot terminate TLS or where OAuth issuer URL cannot be set correctly behind a reverse proxy

Interface

REST API

GraphQL

gRPC

MCP Server

Yes

SDK

Webhooks

Authentication

Methods: OAuth2.1 / OAuth2 using MCP_CLIENT_ID and MCP_CLIENT_SECRET Bearer/client-token style for upstream OpenClaw gateway via OPENCLAW_GATEWAY_TOKEN and per-instance tokens

OAuth: Yes Scopes: No

The README describes an MCP server protected by OAuth2 (AUTH_ENABLED=true, MCP_CLIENT_ID/SECRET, optional MCP_ISSUER_URL). Separate upstream OpenClaw gateway token(s) are used to authenticate to OpenClaw.

Pricing

Free tier: No

Requires CC: No

No pricing information provided; appears self-hosted (Docker/npm).

Agent Metadata

Pagination

none

Idempotent

False

Retry Guidance

Not documented

Known Gotchas

⚠ When behind a reverse proxy, MCP_ISSUER_URL (or --issuer-url) must be set to a public HTTPS URL to avoid OAuth metadata pointing to localhost.
⚠ Pin Docker image versions for production; using :latest is discouraged in the README.
⚠ Correctly configure CORS_ORIGINS (default examples include https://claude.ai).

Alternatives

MCP client integrations that directly call OpenClaw over its own API (if available) Custom proxy/bridge service exposing OpenClaw functionality to Claude in a bespoke way Other MCP servers that provide generic OpenAI/OpenTelemetry tooling for self-hosted agents (if compatible with your backend)

Full Evaluation Report

Comprehensive deep-dive: security analysis, reliability audit, agent experience review, cost modeling, competitive positioning, and improvement roadmap for openclaw-mcp.

AI-powered analysis · PDF + markdown · Delivered within 30 minutes

$99

Package Brief

Quick verdict, integration guide, cost projections, gotchas with workarounds, and alternatives comparison.

Delivered within 10 minutes

Score Monitoring

Get alerted when this package's AF, security, or reliability scores change significantly. Stay ahead of regressions.

Continuous monitoring

$3/mo

API endpoint ↗ Agent guide ↗ Report inaccuracy

Scores are editorial opinions as of 2026-03-30.