Migrate from Firecrawl

Prefetch’s web data endpoints work much like Firecrawl’s v2 API, so most migrations are a change of base URL, header and parameter spelling. This page maps one to the other.

The basics

	Firecrawl v2	Prefetch
Base URL	`https://api.firecrawl.dev/v2`	`https://api.prefetch.io`
Auth header	`Authorization: Bearer fc-…`	`X-API-Key: …`
Parameter style	camelCase	snake_case
Response envelope	`{ success, data }`	`{ success, data, meta }`, and errors add `error_code` and `retryable`
Cached pages	Served up to 2 days old by default (`maxAge`)	Fetched live unless you set `max_age`

To keep Firecrawl’s caching behavior, send max_age=172800000 (2 days, in milliseconds).

Endpoints

Firecrawl v2	Prefetch
`POST /v2/scrape`	`GET` or `POST /scrape`
`POST /v2/map`	`GET /map`
`POST /v2/crawl`	`POST /crawl`
`GET /v2/crawl/{id}`	`GET /crawl/{id}`
`DELETE /v2/crawl/{id}`	`DELETE /crawl/{id}`
`GET /v2/crawl/{id}/errors`	`GET /crawl/{id}/errors`
`POST /v2/batch/scrape`	`POST /batch/scrape`
`GET /v2/batch/scrape/{id}`	`GET /batch/scrape/{id}`
`POST /v2/search`	`GET` or `POST /search`
`/v2/monitor` and its sub-paths	`/monitor` and the same sub-paths

Prefetch also has company data endpoints with no Firecrawl equivalent: /enrich, /brand, /company, /classify, /styleguide, /resolve, /transaction and the logo CDN.

Scrape

Firecrawl	Prefetch	Notes
`url`	`url`
`formats: ["markdown", "html", "rawHtml", "links", "screenshot", "summary"]`	`formats: ["markdown", "html", "raw_html", "links", "screenshot", "summary"]`	Prefetch also has `images`. On `GET`, pass a comma-separated list.
`formats: [{ "type": "json", "prompt": "…", "schema": {…} }]`	`formats: ["json"]` with `json_prompt` and `json_schema`
`onlyMainContent`	`only_main_content`	Both default to `true`.
`includeTags`, `excludeTags`	`include_selectors`, `exclude_selectors`	CSS selectors.
`waitFor`	`wait_for`	Milliseconds. Prefetch also has `wait_for_selector`.
`maxAge`	`max_age`	Milliseconds. Prefetch defaults to a live fetch.
`mobile`	`mobile`
`headers`	`headers`	Requests with headers are never cached on Prefetch.
`removeBase64Images` (default `true`)	`include_base64_images` (default `false`)	Inverted name, same default behavior.
`checkPromptInjection`	`check_prompt_injection`

Map

Firecrawl	Prefetch
`url`, `search`, `limit`	`url`, `search`, `limit`
`sitemap: "include" \| "only" \| "skip"`	`sitemap: "include" \| "only" \| "skip"`
`includeSubdomains`	`include_subdomains`
`ignoreQueryParameters`	`ignore_query_params`

Crawl

Firecrawl	Prefetch	Notes
`limit`	`limit`	Prefetch caps a crawl at 500 pages.
`maxDiscoveryDepth`	`max_depth`	Sitemap pages are depth 0 on both.
`includePaths`, `excludePaths`	`include_paths`, `exclude_paths`	Regular expressions, matched against the path and query on Prefetch.
`allowSubdomains`, `allowExternalLinks`, `crawlEntireDomain`	`allow_subdomains`, `allow_external_links`, `crawl_entire_domain`
`sitemap: "include"` or `"skip"`	`use_sitemap: true` or `false`	For `"only"`, send `use_sitemap: true` with `max_depth: 0`.
`ignoreRobotsTxt`	`respect_robots_txt: false`
`delay` (seconds)	`delay_ms` (milliseconds)
`maxConcurrency`	`concurrency`	1 to 5 on Prefetch.
`scrapeOptions`	`scrape_options`	Takes everything `/scrape` takes.
`webhook`	`webhook`	See Webhooks below.

Batch scrape

Firecrawl takes scrape options at the top level of a batch. Prefetch nests them in scrape_options:

{
  "urls": ["https://example.com/a", "https://example.com/b"],
  "scrape_options": { "formats": ["markdown"], "only_main_content": true },
  "ignore_invalid_urls": true,
  "concurrency": 3
}

ignoreInvalidURLs becomes ignore_invalid_urls, and maxConcurrency becomes concurrency.

Search

Firecrawl	Prefetch
`query`, `limit`, `location`, `country`	`query`, `limit`, `location`, `country`
`sources: ["web"]` or `["news"]`	`type: "web"` or `"news"`
`tbs: "qdr:w"`	`time_range: "week"` (`hour`, `day`, `week`, `month`, `year`)
`includeDomains`, `excludeDomains`	`include_domains`, `exclude_domains`
`scrapeOptions`	`scrape_options`

Reading responses

What	Firecrawl	Prefetch
Scraped markdown	`data.markdown` (a string)	`data.markdown.content`
Source URL	`data.metadata.sourceURL`	`data.metadata.source_url`
Status code	`data.metadata.statusCode`	`data.status_code`
Search results	`data.web[]`	`data.results[]`, each with `scrape` when you asked for content
Crawl pages	`data[]` on the status response	`data.data[]`, with `next` for the following page

Each Prefetch format is an object with counts and flags, such as markdown.word_count and markdown.truncated. See Scrape.

Errors

Both return success: false with an error message. Prefetch adds a stable error_code and a retryable flag, so you can branch on them instead of parsing the message. See Error handling.

Webhooks

Crawl and batch event names carry over: crawl.started, crawl.page, crawl.completed, and batch_scrape.*. Signatures differ:

	Firecrawl	Prefetch
Header	`X-Firecrawl-Signature`	`X-Prefetch-Signature: t=<unix seconds>,v1=<hex>`
What’s signed	The raw body	`<t>.<raw body>`, so you can reject replays older than five minutes
Secret	One per account, in your settings	One per job or monitor, returned when you create it

See Verify signatures for code.

Checklist

Swap the base URL and the auth header.
Rename parameters to snake_case, and nest batch scrape options in scrape_options.
Read data.markdown.content instead of data.markdown.
Set max_age wherever you relied on Firecrawl’s default caching.
Update your webhook signature check.
Branch on error_code and retryable.

No-code tools

Compare

Migrate

The basics

Endpoints

Scrape

Map

Crawl

Batch scrape

Search

Reading responses

Errors

Webhooks

Checklist

​The basics

​Endpoints

​Scrape

​Map

​Crawl

​Batch scrape

​Search

​Reading responses

​Errors

​Webhooks

​Checklist

The basics

Endpoints

Scrape

Map

Crawl

Batch scrape

Search

Reading responses

Errors

Webhooks

Checklist