Multi-agent browser MCP runtime for isolated, persistent, parallel Playwright sessions. Run multiple AI agents without shared tabs or browser state.
Local browser execution for MCP clients, with isolated sessions and explicit page addressing.
- Documentation - Getting started - MCP tool reference
BrowserMesh lets Claude Code, Codex, Cursor, Qwen, and other MCP-compatible clients operate independent Chromium sessions through one local MCP server. Each ready session owns a separate BrowserContext; every page operation names its sessionId and pageId.
The external client reasons and plans. BrowserMesh executes browser operations, enforces isolation, orders work within each session, and returns structured results.
Example: test checkout as a customer while a separate admin session verifies the order.
Install the Playwright-managed Chromium build used by BrowserMesh:
Add BrowserMesh to your MCP client configuration:
The exact configuration shape depends on the client. BrowserMesh reports its installed package version in serverInfo.version during MCP initialization.
Check a local installation without starting the MCP transport:
Chromium and BrowserMesh remain on the user's machine. No hosted BrowserMesh service is required. If Chromium is missing, MCP discovery remains available and browsersessioncreate returns an actionable BROWSERERROR.
BrowserMesh v0.1 is intentionally small: one local Node.js process, one Chromium process, and one non-persistent BrowserContext for every ready session.
- explicit session and page addressing; - isolated browser contexts and page ownership checks; - navigation, inspection, interaction, waits, capture, and persistence; - per-session serialization with parallel execution across indep
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/scrollDynasty--multi-agent-browser-mcp.html)
Hand a whole browser task off in one call: a decision model drives the page server-side, so a flow costs one call, not a turn per click. Ref-based element tables, code-checked assertions, zero-model m
tiliondev/waystoneBrowser action recorder for automation: record a task once, replay it, or hand the folder to an agent
pairbrowse/pairbrowseMultiplayer browser sessions for developers and AI coding agents
693696817/mcd-roundtable麦门圆桌（mcd-roundtable）｜五个 AI 人格辩论，逼出最优麦当劳订单。基于麦当劳 MCP（mcd-mcp / Model Context Protocol）的多智能体点餐决策 Skill：菜单、券、营养、价格全部来自官方接口，金额只认 calculate-price，绝不估算。凑单最省 / 优惠券最大化 / 营养配餐 / 一键下单。Multi-agent McDonald's MCP
christiangrey922/multi-agent-workflow-labTesting and observability for multi-agent delegation, MCP tools, permissions, sandboxed actions, prompts, and workflow replay.
roshanarnav25-sloth/google-flow-mcpMCP server that drives Google Flow (Veo / Nano Banana) through its internal tRPC API, with a hard credit-cost gate
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.