# firecrawl/firecrawl

Source: [GitHub](https://github.com/firecrawl/firecrawl)  
Feed7 permalink: https://feed7.dev/p/firecrawl-1utf6mi  
Published: Unknown  
Trust: Needs Review (needs_review)

## Why Included

Firecrawl Local runs scraping, crawling, and mapping on one machine and can expose a local API or MCP server to coding agents. Some failed local scrapes can fall back to Firecrawl Cloud unless that route is disabled.

## Source Summary

**Firecrawl Local** runs scraping, crawling, and mapping on your machine without a database, queue, or container. It can connect to coding agents as an MCP server and serve the Firecrawl v2 API on localhost; its API routes need no key.

## Practical Implication

For an agent that needs local web context, the CLI can configure Claude Code, Cursor, Windsurf, Gemini CLI, and Codex, backing up config files first. JavaScript-heavy pages need local Chrome or Chromium, and MCP requires **Node.js 22 or newer**.

## Agent-Ready Context

**Firecrawl Local** runs scraping, crawling, and mapping on your machine without a database, queue, or container. It can connect to coding agents as an MCP server and serve the Firecrawl v2 API on localhost; its API routes need no key.

For an agent that needs local web context, the CLI can configure Claude Code, Cursor, Windsurf, Gemini CLI, and Codex, backing up config files first. JavaScript-heavy pages need local Chrome or Chromium, and MCP requires **Node.js 22 or newer**.

Some single-page scrapes that fail locally are fetched through Firecrawl Cloud by default; set `FIRECRAWL_CLOUD=off` to keep work local. Cloud features such as Alexandria are not included in self-hosting, and the project is primarily AGPL-3.0 licensed.

## Connected Context

Feed7 judgment across 1135 accumulated Signals:

Firecrawl Local makes web retrieval a deployable local MCP/API capability, with a specific escape hatch: some failed single-page scrapes still use Cloud unless explicitly disabled. That boundary, plus browser and Node requirements and the AGPL license, makes locality a configuration and deployment property rather than an unconditional guarantee.

- [Panniantong/Agent-Reach](https://feed7.dev/p/agent-reach-0huqxvi) — Agent Reach configures and routes among retrieval backends, while Firecrawl Local supplies one concrete crawler backend; its default Cloud fallback is a relevant constraint when routing is intended to stay local.
- [unclecode/crawl4ai](https://feed7.dev/p/crawl4ai-1cm67rn) — Both provide browser-oriented retrieval for agent pipelines; Firecrawl highlights local operation with a possible Cloud fallback, while Crawl4AI’s supplied account emphasizes keeping the backend current and containing untrusted requests.

## Context Map

- Layer: tools
- Domains: coding, research
- Topics: mcp, retrieval, coding-agents

## Uncertainty

- Some single-page scrapes that fail locally are fetched through Firecrawl Cloud by default; set `FIRECRAWL_CLOUD=off` to keep work local. Cloud features such as Alexandria are not included in self-hosting, and the project is primarily AGPL-3.0 licensed.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.
