Firecrawl

Firecrawl turns any website into LLM-ready data: scrape single pages to markdown or JSON, crawl entire sites, map all URLs, search the web with full page content inlined, or run autonomous agents and scripted browser sessions. Best when you need clean structured content from a URL or site — unlike search APIs (Serper, Brave, Exa) that return result links, Firecrawl renders pages (JS included) and extracts readable content. Unlike Cloudflare Browser Rendering, Firecrawl adds multi-page crawl orchestration, sitemap-aware mapping, AI-powered structured extraction, and an agent runtime. 23 endpoints available through Lava’s AI Gateway. See the Firecrawl API docs for full documentation.

Supports both managed (Lava’s API keys) and unmanaged (bring your own credentials) mode.

Endpoints

Scrape a single URL and return clean markdown and/or structured JSON. Supports JavaScript rendering, main-content extraction, and multiple output formats (markdown, html, rawHtml, links, screenshot, screenshot@fullPage, json). 1 credit per page.

POST https://api.firecrawl.dev/v2/scrape — $0.001 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/scrape', { body: {"url":"https://example.com","formats":["markdown"]} });

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fscrape" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com","formats":["markdown"]}'

Get a complete list of URLs from a website fast. Use this before crawling to scope a site, or as a lightweight sitemap discovery. 1 credit.

POST https://api.firecrawl.dev/v2/map — $0.001 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/map', { body: {"url":"https://example.com","limit":100} });

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fmap" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com","limit":100}'

Search the web and return full page content inline in markdown. Combines a search query with page scraping in one call. 2 credits per 10 results.

POST https://api.firecrawl.dev/v2/search — $0.002 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/search', {
  body: {
"query": "latest AI agent frameworks",
"limit": 10,
"scrapeOptions": {
  "formats": [
    "markdown"
  ]
}
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fsearch" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query":"latest AI agent frameworks","limit":10,"scrapeOptions":{"formats":["markdown"]}}'

Kick off a recursive site crawl — returns a job id. Poll /v2/crawl/{id} for results. Billed 1 credit per page crawled (not captured at kick-off; heavy crawls under-collect at the gateway).

POST https://api.firecrawl.dev/v2/crawl — $0.001 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/crawl', {
  body: {
"url": "https://example.com",
"limit": 25,
"scrapeOptions": {
  "formats": [
    "markdown"
  ]
}
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fcrawl" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com","limit":25,"scrapeOptions":{"formats":["markdown"]}}'

List currently active crawl jobs. Free.

GET https://api.firecrawl.dev/v2/crawl/active — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/crawl/active', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fcrawl%2Factive" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Get status and results of a crawl job. Free status poll.

GET https://api.firecrawl.dev/v2/crawl/550e8400-e29b-41d4-a716-446655440000 — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/crawl/550e8400-e29b-41d4-a716-446655440000', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fcrawl%2F550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Get errors from a crawl job. Free.

GET https://api.firecrawl.dev/v2/crawl/550e8400-e29b-41d4-a716-446655440000/errors — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/crawl/550e8400-e29b-41d4-a716-446655440000/errors', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fcrawl%2F550e8400-e29b-41d4-a716-446655440000%2Ferrors" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Cancel an in-progress crawl job. Free.

DELETE https://api.firecrawl.dev/v2/crawl/550e8400-e29b-41d4-a716-446655440000 — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/crawl/550e8400-e29b-41d4-a716-446655440000', { method: 'DELETE' });

curl -X DELETE "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fcrawl%2F550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json"

Preview the params Firecrawl will derive for a crawl from a natural-language prompt, without kicking a job off. Free.

POST https://api.firecrawl.dev/v2/crawl/params-preview — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/crawl/params-preview', {
  body: {
"url": "https://example.com",
"prompt": "Extract pricing and documentation pages"
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fcrawl%2Fparams-preview" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://example.com","prompt":"Extract pricing and documentation pages"}'

Submit a list of URLs to scrape as a single job. Returns a job id — poll /v2/batch/scrape/{id}. Billed 1 credit per URL (not captured at kick-off).

POST https://api.firecrawl.dev/v2/batch/scrape — $0.001 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/batch/scrape', {
  body: {
"urls": [
  "https://example.com",
  "https://example.org"
],
"formats": [
  "markdown"
]
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fbatch%2Fscrape" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"urls":["https://example.com","https://example.org"],"formats":["markdown"]}'

Get status and results of a batch scrape job. Free.

GET https://api.firecrawl.dev/v2/batch/scrape/550e8400-e29b-41d4-a716-446655440000 — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/batch/scrape/550e8400-e29b-41d4-a716-446655440000', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fbatch%2Fscrape%2F550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Get errors from a batch scrape job. Free.

GET https://api.firecrawl.dev/v2/batch/scrape/550e8400-e29b-41d4-a716-446655440000/errors — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/batch/scrape/550e8400-e29b-41d4-a716-446655440000/errors', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fbatch%2Fscrape%2F550e8400-e29b-41d4-a716-446655440000%2Ferrors" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Cancel an in-progress batch scrape job. Free.

DELETE https://api.firecrawl.dev/v2/batch/scrape/550e8400-e29b-41d4-a716-446655440000 — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/batch/scrape/550e8400-e29b-41d4-a716-446655440000', { method: 'DELETE' });

curl -X DELETE "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fbatch%2Fscrape%2F550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json"

Extract structured data from one or more URLs against a prompt or schema. Returns a job id — poll /v2/extract/{id}. Upstream pricing is token-based; flat kick-off cost under-collects on heavy extractions.

POST https://api.firecrawl.dev/v2/extract — $0.001 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/extract', {
  body: {
"urls": [
  "https://example.com"
],
"prompt": "Extract the title and main heading"
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fextract" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"urls":["https://example.com"],"prompt":"Extract the title and main heading"}'

Get status and results of an extract job. Free.

GET https://api.firecrawl.dev/v2/extract/550e8400-e29b-41d4-a716-446655440000 — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/extract/550e8400-e29b-41d4-a716-446655440000', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fextract%2F550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Create a remote browser session for scripted interaction. Upstream bills 2 credits/minute; gateway bills flat kick-off and under-collects on long sessions.

POST https://api.firecrawl.dev/v2/browser — $0.001 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/browser', { body: {} });

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fbrowser" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{}'

Execute JavaScript in a live browser session. 5 credits per action upstream.

POST https://api.firecrawl.dev/v2/browser/550e8400-e29b-41d4-a716-446655440000/execute — $0.005 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/browser/550e8400-e29b-41d4-a716-446655440000/execute', { body: {"code":"return document.title;"} });

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fbrowser%2F550e8400-e29b-41d4-a716-446655440000%2Fexecute" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"code":"return document.title;"}'

List active browser sessions. Free.

GET https://api.firecrawl.dev/v2/browser — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/browser', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fbrowser" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

End a browser session. Free.

DELETE https://api.firecrawl.dev/v2/browser/550e8400-e29b-41d4-a716-446655440000 — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/browser/550e8400-e29b-41d4-a716-446655440000', { method: 'DELETE' });

curl -X DELETE "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fbrowser%2F550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json"

Execute JavaScript or run an AI agent prompt inside a live browser session bound to a prior scrape. Provide `code` OR `prompt`, never both. 5 credits per action upstream.

POST https://api.firecrawl.dev/v2/scrape/550e8400-e29b-41d4-a716-446655440000/interact — $0.005 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/scrape/550e8400-e29b-41d4-a716-446655440000/interact', {
  body: {
"code": "const title = await page.title(); JSON.stringify({ title });",
"language": "node"
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fscrape%2F550e8400-e29b-41d4-a716-446655440000%2Finteract" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"code":"const title = await page.title(); JSON.stringify({ title });","language":"node"}'

Run an autonomous web-gathering agent against a goal. Returns a job id. Upstream pricing is dynamic (5 free daily runs, then variable); flat kick-off cost under-collects on long runs.

POST https://api.firecrawl.dev/v2/agent — $0.001 / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/agent', {
  body: {
"prompt": "Find the pricing page on https://example.com and return the lowest paid tier"
},
});

curl -X POST "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fagent" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json" \
  -d '{"prompt":"Find the pricing page on https://example.com and return the lowest paid tier"}'

Get status and output of an agent run. Free.

GET https://api.firecrawl.dev/v2/agent/550e8400-e29b-41d4-a716-446655440000 — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/agent/550e8400-e29b-41d4-a716-446655440000', { method: 'GET' });

curl "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fagent%2F550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY"

Cancel an in-progress agent run. Free.

DELETE https://api.firecrawl.dev/v2/agent/550e8400-e29b-41d4-a716-446655440000 — Free / request

SDK
cURL

const data = await lava.gateway('https://api.firecrawl.dev/v2/agent/550e8400-e29b-41d4-a716-446655440000', { method: 'DELETE' });

curl -X DELETE "https://api.lava.so/v1/forward?u=https%3A%2F%2Fapi.firecrawl.dev%2Fv2%2Fagent%2F550e8400-e29b-41d4-a716-446655440000" \
  -H "Authorization: Bearer $LAVA_SECRET_KEY" \
  -H "Content-Type: application/json"

Next Steps

All Providers

Browse all supported AI providers

Forward Proxy

Learn how to construct proxy URLs and authenticate requests

⌘I

Get Started

AI Gateway

AI Spend

Monetize

Integration

SDK

Cookbook

Comparisons

Endpoints

Scrape a single URL and return clean markdown and/or structured JSON. Supports JavaScript rendering, main-content extraction, and multiple output formats (markdown, html, rawHtml, links, screenshot, screenshot@fullPage, json). 1 credit per page.

Get a complete list of URLs from a website fast. Use this before crawling to scope a site, or as a lightweight sitemap discovery. 1 credit.

Search the web and return full page content inline in markdown. Combines a search query with page scraping in one call. 2 credits per 10 results.

Kick off a recursive site crawl — returns a job id. Poll /v2/crawl/{id} for results. Billed 1 credit per page crawled (not captured at kick-off; heavy crawls under-collect at the gateway).

List currently active crawl jobs. Free.

Get status and results of a crawl job. Free status poll.

Get errors from a crawl job. Free.

Cancel an in-progress crawl job. Free.

Preview the params Firecrawl will derive for a crawl from a natural-language prompt, without kicking a job off. Free.

Submit a list of URLs to scrape as a single job. Returns a job id — poll /v2/batch/scrape/{id}. Billed 1 credit per URL (not captured at kick-off).

Get status and results of a batch scrape job. Free.

Get errors from a batch scrape job. Free.

Cancel an in-progress batch scrape job. Free.

Extract structured data from one or more URLs against a prompt or schema. Returns a job id — poll /v2/extract/{id}. Upstream pricing is token-based; flat kick-off cost under-collects on heavy extractions.

Get status and results of an extract job. Free.

Create a remote browser session for scripted interaction. Upstream bills 2 credits/minute; gateway bills flat kick-off and under-collects on long sessions.

Execute JavaScript in a live browser session. 5 credits per action upstream.

List active browser sessions. Free.

End a browser session. Free.

Execute JavaScript or run an AI agent prompt inside a live browser session bound to a prior scrape. Provide `code` OR `prompt`, never both. 5 credits per action upstream.

Run an autonomous web-gathering agent against a goal. Returns a job id. Upstream pricing is dynamic (5 free daily runs, then variable); flat kick-off cost under-collects on long runs.

Get status and output of an agent run. Free.

Cancel an in-progress agent run. Free.

Next Steps

All Providers

Forward Proxy

​Endpoints

​Scrape a single URL and return clean markdown and/or structured JSON. Supports JavaScript rendering, main-content extraction, and multiple output formats (markdown, html, rawHtml, links, screenshot, screenshot@fullPage, json). 1 credit per page.

​Get a complete list of URLs from a website fast. Use this before crawling to scope a site, or as a lightweight sitemap discovery. 1 credit.

​Search the web and return full page content inline in markdown. Combines a search query with page scraping in one call. 2 credits per 10 results.

​Kick off a recursive site crawl — returns a job id. Poll /v2/crawl/{id} for results. Billed 1 credit per page crawled (not captured at kick-off; heavy crawls under-collect at the gateway).

​List currently active crawl jobs. Free.

​Get status and results of a crawl job. Free status poll.

​Get errors from a crawl job. Free.

​Cancel an in-progress crawl job. Free.

​Preview the params Firecrawl will derive for a crawl from a natural-language prompt, without kicking a job off. Free.

​Submit a list of URLs to scrape as a single job. Returns a job id — poll /v2/batch/scrape/{id}. Billed 1 credit per URL (not captured at kick-off).

​Get status and results of a batch scrape job. Free.

​Get errors from a batch scrape job. Free.

​Cancel an in-progress batch scrape job. Free.

​Extract structured data from one or more URLs against a prompt or schema. Returns a job id — poll /v2/extract/{id}. Upstream pricing is token-based; flat kick-off cost under-collects on heavy extractions.

​Get status and results of an extract job. Free.

​Create a remote browser session for scripted interaction. Upstream bills 2 credits/minute; gateway bills flat kick-off and under-collects on long sessions.

​Execute JavaScript in a live browser session. 5 credits per action upstream.

​List active browser sessions. Free.

​End a browser session. Free.

​Execute JavaScript or run an AI agent prompt inside a live browser session bound to a prior scrape. Provide code OR prompt, never both. 5 credits per action upstream.

​Run an autonomous web-gathering agent against a goal. Returns a job id. Upstream pricing is dynamic (5 free daily runs, then variable); flat kick-off cost under-collects on long runs.

​Get status and output of an agent run. Free.

​Cancel an in-progress agent run. Free.

​Next Steps

All Providers

Forward Proxy

Endpoints

Scrape a single URL and return clean markdown and/or structured JSON. Supports JavaScript rendering, main-content extraction, and multiple output formats (markdown, html, rawHtml, links, screenshot, screenshot@fullPage, json). 1 credit per page.

Get a complete list of URLs from a website fast. Use this before crawling to scope a site, or as a lightweight sitemap discovery. 1 credit.

Search the web and return full page content inline in markdown. Combines a search query with page scraping in one call. 2 credits per 10 results.

Kick off a recursive site crawl — returns a job id. Poll /v2/crawl/{id} for results. Billed 1 credit per page crawled (not captured at kick-off; heavy crawls under-collect at the gateway).

List currently active crawl jobs. Free.

Get status and results of a crawl job. Free status poll.

Get errors from a crawl job. Free.

Cancel an in-progress crawl job. Free.

Preview the params Firecrawl will derive for a crawl from a natural-language prompt, without kicking a job off. Free.

Submit a list of URLs to scrape as a single job. Returns a job id — poll /v2/batch/scrape/{id}. Billed 1 credit per URL (not captured at kick-off).

Get status and results of a batch scrape job. Free.

Get errors from a batch scrape job. Free.

Cancel an in-progress batch scrape job. Free.

Extract structured data from one or more URLs against a prompt or schema. Returns a job id — poll /v2/extract/{id}. Upstream pricing is token-based; flat kick-off cost under-collects on heavy extractions.

Get status and results of an extract job. Free.

Create a remote browser session for scripted interaction. Upstream bills 2 credits/minute; gateway bills flat kick-off and under-collects on long sessions.

Execute JavaScript in a live browser session. 5 credits per action upstream.

List active browser sessions. Free.

End a browser session. Free.

Execute JavaScript or run an AI agent prompt inside a live browser session bound to a prior scrape. Provide `code` OR `prompt`, never both. 5 credits per action upstream.

Run an autonomous web-gathering agent against a goal. Returns a job id. Upstream pricing is dynamic (5 free daily runs, then variable); flat kick-off cost under-collects on long runs.

Get status and output of an agent run. Free.

Cancel an in-progress agent run. Free.

Next Steps