mcp-browser-use

Provides an MCP server that connects an MCP client (e.g., Claude/Cursor) to the browser automation capabilities of browser-use, allowing agents to control a browser through the MCP client while reusing the LLM already configured in that MCP client.

Evaluated Mar 30, 2026 (22d ago)
Repo ↗ DevTools mcp browser-automation playwright agentic python fastapi uvicorn
⚙ Agent Friendliness
32
/ 100
Can an agent use this?
🔒 Security
33
/ 100
Is it safe for agents?
⚡ Reliability
30
/ 100
Does it work consistently?

Score Breakdown

⚙ Agent Friendliness

MCP Quality
55
Documentation
40
Error Messages
0
Auth Simplicity
30
Rate Limits
0

🔒 Security

TLS Enforcement
60
Auth Strength
20
Scope Granularity
0
Dep. Hygiene
55
Secret Handling
40

No explicit security guidance is documented in the provided README/manifest. The service likely runs a local MCP server and uses Playwright/browser-use, which increases exposure to untrusted web content and potential data leakage through browsing context. TLS/auth/rate-limit and secret-handling practices are not described; treat defaults as unknown and secure the host/container and network accordingly.

⚡ Reliability

Uptime/SLA
0
Version Stability
45
Breaking Changes
50
Error Recovery
25
AF Security Reliability

Best When

You already have an MCP client configured with the desired LLM and want browser-use capabilities exposed to it via MCP.

Avoid When

You cannot ensure the MCP client/host environment is secured, or you need clear guarantees about authentication, rate limiting, and data handling for this specific package.

Use Cases

  • Agent-driven web browsing and task completion via MCP clients
  • Automating repetitive browser workflows (search, navigation, form filling)
  • Building agent tooling that needs browser context without directly embedding a browser automation stack in the agent

Not For

  • Use as a general-purpose public web API without an MCP client (no evidence of hardened multi-tenant exposure)
  • High-security, audited environments requiring strict authn/z guarantees specific to this service (insufficient documented details)
  • Environments where outbound browser automation or Playwright-driven behavior is prohibited

Interface

REST API
No
GraphQL
No
gRPC
No
MCP Server
Yes
SDK
No
Webhooks
No

Authentication

OAuth: No Scopes: No

The README does not document any auth mechanism for this MCP server. It implies using the LLM already set up in the MCP client, but does not describe server-side access control.

Pricing

Free tier: No
Requires CC: No

No pricing information provided; README mentions avoiding paying for an additional LLM API key because it uses the MCP client's already-configured LLM.

Agent Metadata

Pagination
none
Idempotent
False
Retry Guidance
Not documented

Known Gotchas

  • Browser automation can be nondeterministic (timing, selectors, page changes). Agents may need robust retry/backoff at the task level.
  • Because auth/rate-limit/error semantics are not documented here, agent frameworks may need defensive handling (timeouts, reconnects, and conservative concurrency).
  • The README does not specify required environment variables or how the MCP client passes LLM/browser settings; misconfiguration may lead to startup/runtime failures.

Alternatives

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

$3

Score Monitoring

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

Continuous monitoring

$3/mo

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

8642
Packages Evaluated
17761
Need Evaluation
586
Need Re-evaluation
Community Powered