mcp-browser-use

Provides an MCP server (built on FastMCP/FastAPI) that exposes a single tool `run_browser_agent` to orchestrate a `browser-use` automated browser session, using LLM providers for browser actions and returning the final result to the MCP client.

Evaluated Mar 30, 2026 (21d ago)

Repo ↗ Automation mcp fastmcp browser-automation fastapi agentic-ai python chromium browser-use

⚙ 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

Server controls a full browser instance, which can access external sites and system resources depending on how it’s deployed. The README recommends reviewing SECURITY.md and keeping credentials in environment variables, but it does not document authentication/authorization for MCP tool invocation. TLS/authZ specifics and rate/egress restrictions are not evidenced in the provided text.

⚡ Reliability

Uptime/SLA

Version Stability

Breaking Changes

Error Recovery

Best When

You need agent-based web automation accessible through MCP and can provide/secure the required LLM API keys and browser environment configuration.

Avoid When

You cannot control where the browser navigates, who can invoke the MCP tool, or how outbound network access is restricted.

Use Cases

• Automated web navigation and task completion driven by natural-language instructions
• Integrating browser automation into Claude Desktop or other MCP-capable clients
• End-to-end browsing workflows (forms, clicks, reading content, screenshots) within an agent loop

Not For

• Running untrusted browsing targets or exposing the server to untrusted clients without strong isolation
• Use cases requiring strict determinism or guaranteed UI stability
• High-volume production automation without verifying rate limiting and resource controls

Interface

REST API

GraphQL

gRPC

MCP Server

Yes

SDK

Webhooks

Authentication

Methods: Environment-variable API keys for multiple LLM providers (e.g., ANTHROPIC_API_KEY, OPENAI_API_KEY, etc.)

OAuth: No Scopes: No

The README describes provider API keys and model settings via environment variables; it does not document authentication/authorization for who can call the MCP server tool.

Pricing

Free tier: No

Requires CC: No

No pricing information for the server itself; costs depend on the selected LLM provider usage and the volume/complexity of browsing steps.

Agent Metadata

Pagination

none

Idempotent

False

Retry Guidance

Not documented

Known Gotchas

⚠ Browser automation is sensitive to UI changes, dynamic pages, and slow loads; agent runs may fail or take variable time.
⚠ LLM/provider configuration and rate/step limits (e.g., max steps/actions) strongly affect outcomes.
⚠ Remote browser control can introduce security risks; ensure you follow the repository’s SECURITY guidance and restrict invocation sources.

Alternatives

browser-use used directly from your own app Other MCP/agent browser automation servers Playwright/Puppeteer scripts with an LLM planner (separate orchestration)

Full Evaluation Report

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

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.