Firecrawl API migration
Evaluate the documented Firecrawl-shaped REST subset in AIHackr. Compare supported routes, MCP tools and differences before changing your integration.
On this page
A migration surface to evaluate
AIHackr exposes a documented Firecrawl-shaped REST subset. It uses Cloudflare Browser Run and official Cloudflare OSS MCP tools; it does not run the Firecrawl engine or call Firecrawl Cloud. This is not a guarantee that a Firecrawl SDK or every request option will work unchanged.
Supported routes
| Route | AIHackr behavior |
|---|---|
| POST /v2/scrape | One public URL, exactly one output format: markdown, html or links. |
| POST /v2/map | Same-origin links found on one submitted page. Not a full-site index. |
| POST /v2/crawl | Bounded asynchronous same-origin crawl. |
| GET /v2/crawl/:id | Workspace-owned status, retained records and relative pagination URL. |
| DELETE /v2/crawl/:id | Cancel a workspace-owned crawl. |
| POST /v2/extract | One URL, a prompt and/or JSON schema. |
| Search and batch scrape | Unavailable. Unsupported requests return explicit errors. |
Migration checklist
- Create an AIHackr account and use an AIHackr customer key.
- Change the base URL to https://aihackr.com and test direct HTTP requests first.
- Check every request field against the relevant reference. Unknown options are rejected.
- Update response parsing and crawl pagination. Preserve metadata and page outcomes.
- Set explicit, small crawl limits. Trial accounting reserves the requested maximum.
- Switch MCP tool names to the Cloudflare tool list if using agents.
- Review authentication, retention and failure behavior before sending a larger workload.
Differences that matter
AIHackr currently has no web search, batch scraping, target cookies, arbitrary headers, browser actions or guaranteed anti-bot bypass. MCP names follow Cloudflare’s server rather than Firecrawl’s. Paid billing and hosted OAuth are not live. There is no published benchmark proving performance or feature equivalence.
Start with a single-page request and compare the actual result your application needs. See limits before designing recurring ingestion.