lucasmartins-ai/lookacrawler

Overview

Open-source TypeScript crawler that runs locally as a token-efficient alternative to Firecrawl, with a native MCP server for LLM integration. It reduces token usage by over 73% when feeding crawled web content into language model workflows.

Best for

Best for
Developers who want a local, token-conscious crawler wired into MCP-based AI workflows

Use cases

• Crawl web pages for LLM context while minimizing token consumption
• Expose scraped content to MCP-compatible agents and assistants
• Self-host a Firecrawl-style crawling service without cloud costs

Notes

Open-source TypeScript crawler that runs locally as a token-efficient alternative to Firecrawl, with a native MCP server for LLM integration. It reduces token usage by over 73% when feeding crawled web content into language model workflows.

2 stars on GitHub. Last updated 2026-09-04. Licensed MIT.

Use cases

• Crawl web pages for LLM context while minimizing token consumption
• Expose scraped content to MCP-compatible agents and assistants
• Self-host a Firecrawl-style crawling service without cloud costs

Pros

• Free and self-hosted with no vendor lock-in
• Claims significant token reduction for LLM pipelines
• Native MCP server enables direct integration with MCP clients

Cons

• Very early project with only 2 GitHub stars
• Lacks the maturity, documentation, and community of established crawlers like Firecrawl
• No evidence of production hardening or feature parity with Firecrawl

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.