browsegrab is a lightweight browser automation library designed for local LLMs, combining Playwright's accessibility tree with MarkGrab's HTML-to-markdown conversion for efficient token usage.
From the registry: Token-efficient browser agent for local LLMs. Playwright + accessibility tree + MarkGrab.
$ pip install browsegrab[mcp] && playwright install chromiumPlease install the `browsegrab` MCP server into my current AI client (that's you).
Required prerequisites (do these first if not already done):
- **Install browsegrab[mcp]** — Install the package with MCP extras and Playwright Chromium Run: `pip install browsegrab[mcp] && playwright install chromium`
Canonical MCP server config (stdio transport):
- command: `browsegrab-mcp`
- args: []
- optional environment variables:
- `BROWSEGRAB_BROWSER_HEADLESS`: Run browser headless (example: `true`)
- `BROWSEGRAB_BROWSER_TIMEOUT_MS`: Browser timeout (ms) (example: `30000`)
- `BROWSEGRAB_LLM_PROVIDER`: LLM backend provider (example: `vllm`)
- `BROWSEGRAB_LLM_BASE_URL`: LLM base URL (example: `http://localhost:8000/v1`)
- `BROWSEGRAB_LLM_MODEL`: LLM model ID (example: `Qwen/Qwen3.5-32B-AWQ`)
- `BROWSEGRAB_AGENT_MAX_STEPS`: Agent max steps (example: `10`)
- `BROWSEGRAB_AGENT_ENABLE_CACHE`: Enable success pattern cache (example: `true`)
Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.BROWSEGRAB_LLM_PROVIDERrequiredSpecifies the LLM provider to use (vllm, ollama, or openai).BROWSEGRAB_LLM_BASE_URLrequiredBase URL for the LLM service.BROWSEGRAB_LLM_MODELrequiredSpecifies the model to use for the LLM.MCP server for Firecrawl web scraping, structured data extraction and web search integration.