BROWSER USE

- Browser Use Agents: give Browser Use a task and receive completed work. API V4 is current for new integrations.
- Browser Infrastructure: connect your agent or automation to managed browsers through SDK, REST, or CDP. Starts at $0.02/browser-hour.
- Developer tools: Open Source, Browser Harness, SDK, and MCP support the two products above.

[Developer Index](https://browser-use.com/index.md)
[Product Map](https://browser-use.com/llms.txt)
[Full Product Context](https://browser-use.com/llms-full.txt)
[Pricing](https://browser-use.com/pricing.md)
[Cloud Docs](https://docs.browser-use.com/cloud/quickstart)
[Open Source Docs](https://docs.browser-use.com/open-source/introduction)

---

# Playwright MCP vs Browser Harness

Comparison | [Browser Harness](https://browser-use.com/browser-harness)

## Playwright MCP vs Browser Harness.

Same score on 100 public browser tasks, using half the median input tokens per task. And Browser Harness works in your real Chrome, logged in as you.

[Try it in the cloud](https://cloud.browser-use.com/v4?utm_source=browser-use&utm_medium=playwright-mcp-page)

Connect your agent

[QA my website end to end. Test signup, checkout, mobile layout, and console and network errors. Return screenshots and a prioritized fix list.](https://browser-use.com/showcase/videos/qa-site.mp4)

PromptQA my website end to end. Test signup, checkout, mobile layout, and console and network errors. Return screenshots and a prioritized fix list.

## Side by side

Browser Harness Playwright MCP

Score on 100 browser tasks: Browser Harness: 60.4 / 100; Playwright MCP: 60.0 / 100, a statistical tie

Input tokens per task (median): Browser Harness: 2.1M; Playwright MCP: 4.4M

Agent cost per task (median): Browser Harness: $0.042; Playwright MCP: $0.066

Agent time per task (median): Browser Harness: 9.0 min; Playwright MCP: 8.3 min

Browser: Browser Harness: Your Chrome, with your logins; Playwright MCP: Its own browser profile. Your Chrome needs the Playwright Extension

Setup: Browser Harness: Paste one prompt into your agent; Playwright MCP: Add an MCP server to your agent's config

Missing tool: Browser Harness: The agent writes it and keeps it; Playwright MCP: Fixed tool set, more behind --caps flags

Many browsers at once: Browser Harness: Browser Use cloud, $0.02 per browser-hour; Playwright MCP: Point --cdp-endpoint at a remote browser

License: Browser Harness: MIT; Playwright MCP: Apache-2.0

Pick Playwright MCP for a clean, isolated browser in tests, or if your team already runs Playwright.

Source: [Matched eval: method and per-task scores](https://browser-use.com/benchmarks/coding-agent-browser-tools.json), [BU Bench V2.1 tasks](https://github.com/browser-use/benchmark), [Playwright MCP README](https://github.com/microsoft/playwright-mcp), checked Oct 2026.

## Questions

Is Browser Harness better than Playwright MCP?

On success, neither wins. On 100 public BU Bench V2.1 tasks with the same agent (Codex, GPT-6 Luna), the same cloud browser and the same judge, Browser Harness scored 60.4 and Playwright MCP 60.0 out of 100; the paired difference has a 95% interval of -5.7 to +6.6 points, which includes zero. The difference is cost: Playwright MCP used 2.1x the input tokens per task (median 4.4M vs 2.1M), so the agent's bill was $0.066 vs $0.042 per task. One run per task; per-task scores are public.

Can I use both?

Yes, they do not conflict. Use Playwright MCP for isolated test runs and Browser Harness for work on sites you are logged in to.

Does Browser Harness have an MCP server?

Yes, browser-harness-mcp, for any MCP client. You usually do not need it: the setup prompt installs Browser Harness as a skill, with no config to edit.

What does it cost?

Nothing. Browser Harness is open source under MIT, and using your own Chrome is free. Cloud browsers are $0.02 per browser-hour if you want them.

You will never use  
the browser again.

[Try it in the cloud](https://cloud.browser-use.com/v4?utm_source=browser-use&utm_medium=playwright-mcp-page)

Connect your agent