MCP servers · quartzunit

Browsegrab: browser automation for local LLM agents

A Python browser automation library for local LLM agents, with a built-in MCP server exposing eight browser tools.

This text was machine-translated from the publisher's own information and has not been checked by a human editor. The publisher's original text is the reference.

This resource was imported automatically from the sources listed below and has not been reviewed by the AI Gate team yet. The description is the publisher's own, as published. Check the repository before installing.

What it does

Browsegrab is a browser automation library that combines Playwright's accessibility tree with MarkGrab's HTML-to-markdown conversion. Elements receive stable references such as e1 and e2, so the model does not need a vision model to act on them. It also caches successful workflows, so repeated workflows can run with zero LLM calls, and it includes a 5-stage JSON parser for action outputs. Its built-in MCP server exposes eight tools: navigate, click, type, snapshot, scroll, extract content, go back and wait.

It is intended for agents run by local models, such as those served through vLLM, Ollama or OpenAI-compatible endpoints, where the tokens sent per step matter. The publisher reports about 500 to 1,500 tokens per step, compared with 4,000 to 10,000 for browser-use, and about 60 tokens for a single-element page such as example.com versus about 1,254 for a Wikipedia article with 452 interactive elements.

It can be used through a Python API, a CLI or as an MCP server connected to clients such as Claude Desktop, Cursor or VS Code. Settings are read from environment variables with the BROWSEGRAB_ prefix. It is part of the QuartzUnit ecosystem alongside markgrab, snapgrab and docpick, and is released under the MIT license.

Limitations and requirements

Designed for local LLMs of roughly 8B to 35B parameters. Core dependencies are playwright and httpx. Configuration is done through BROWSEGRAB_* environment variables.

Safety

Our team has not reviewed this resource. Only automated rules checked it; resources with high-risk signals are not published. These are signals, not a guarantee that installing it is safe. Try it in a sandbox project first.

  • Runs code or changes an assistant's behaviour

Sources

  • Official MCP Registry first seen 13 Nov 2026 · last seen 13 Nov 2026 · its page on the source
    Browsegrab

    Token-efficient browser agent for local LLMs. Playwright + accessibility tree + MarkGrab.

    v0.1.1
    As shown on registry.modelcontextprotocol.io when we last read it
MCP servers
cloudflare

Cloudflare MCP servers

Cloudflare's hosted MCP servers for its products: documentation search, sandbox containers, browser rendering, Radar internet data, logs and analytics, Workers and more.

4.5K GitHub starsCloud & DevOpsSoftware development
Updated 17 Nov 2026
MCP servers
redis

Redis MCP Server

Redis's official MCP server that lets an AI assistant store, read and search data in a Redis database, including streams, JSON documents and vector indexes.

642 GitHub starsData & databasesSoftware development
Updated 17 Nov 2026
MCP servers
microsoft

Playwright MCP (browser automation)

Microsoft's MCP server that lets an AI assistant open web pages, click, type and read them through a real browser, using the page structure instead of screenshots.

39.6K GitHub starsTesting & QAWeb, search & browser
Updated 17 Nov 2026
MCP servers
hashicorp

Terraform MCP Server (HashiCorp)

HashiCorp's MCP server that gives an assistant current Terraform Registry information on providers, modules and policies, and can manage HCP Terraform workspaces.

1.6K GitHub starsCloud & DevOpsSoftware development
Updated 17 Nov 2026