Migrate from Firecrawl

FairCrawl is intentionally close to the Firecrawl mental model, but the request shapes are cleaner and the runtime defaults are different.

High-level mapping

Firecrawl	FairCrawl
https://api.firecrawl.dev/v2	https://api.faircompany.ai/v1
formats: ["markdown"]	format: "markdown"
onlyMainContent	only_main_content
includeSubdomains	include_subdomains
sitemap: "only"	sitemap_only: true
async /crawl job start + polling	synchronous POST /v1/crawl/web/crawl
/agent prompt	/agent instruction
/agent schema	/agent json_schema
credits budgeting	budget_usd in USD

Scrape rewrite

 POST https://api.firecrawl.dev/v2/scrape
 POST https://api.faircompany.ai/v1/crawl/web/scrape

{
  "url": "https://stripe.com/pricing",
  "formats": ["markdown"],
  "onlyMainContent": true
  "format": "markdown",
  "only_main_content": true
}

Map rewrite

 POST https://api.firecrawl.dev/v2/map
 POST https://api.faircompany.ai/v1/crawl/map

{
  "url": "https://docs.stripe.com",
  "limit": 100,
  "includeSubdomains": true,
  "sitemap": "only"
  "include_subdomains": true,
  "sitemap_only": true
}

Crawl rewrite

Firecrawl’s crawl flow starts a job and then polls it. FairCrawl’s current /web/crawl path returns the completed page set directly.

 POST https://api.firecrawl.dev/v2/crawl
 POST https://api.faircompany.ai/v1/crawl/web/crawl

{
  "url": "https://docs.python.org/3/index.html",
  "limit": 50,
  "deduplicateSimilarURLs": true,
  "regexOnFullURL": "pricing|billing"
}

Agent rewrite

 POST https://api.firecrawl.dev/v2/agent
 POST https://api.faircompany.ai/v1/crawl/agent

{
  "prompt": "Compare Stripe and Adyen pricing",
  "schema": { "type": "object" },
  "maxCredits": 50
  "instruction": "Compare Stripe and Adyen pricing",
  "response_format": "json",
  "json_schema": { "type": "object" },
  "budget_usd": 0.05
}

Find-and-replace checklist

1. Replace formats with a single format.
2. Convert camelCase body keys to the snake_case fields FairCrawl exposes publicly.
3. Remove Firecrawl crawl polling code and treat FairCrawl /web/crawl as a direct response.
4. Swap agent prompt for instruction and schema schema for json_schema.
5. Keep platform-specific Firecrawl code behind source-filtered /research or direct FairCrawl typed routes, depending on the use case.