Introducing the Firecrawl Developer Index, built for supercharging coding agents. Read the announcement →
2 Months Free - Annually

AI Agents &
Knowledge Bases

Fresh knowledge bases for AI chatbots and RAG products.
Crawl docs, parse PDFs, refresh on change, and cite sources on one API.

//
Used by over 1.25M developers
//
Trusted by 150,000+
companies
of all sizes
10x
faster knowledge base setup
99.9%
content accuracy vs manual uploads
24/7
automatic refresh schedules

Perfect for

Customer support chatbots

Deflect tickets with grounded answers from your docs, FAQs, and troubleshooting guides, with citations back to source pages.

AI agent builders

Keep every agent synced with the latest product docs, help center updates, and changelogs without manual exports or brittle scrapers.

Sales and solutions assistants

Answer product, pricing, and integration questions accurately using up-to-date public pages and technical docs, not stale PDFs.

Internal enablement copilots

Power employee-facing assistants with scoped knowledge bases so teams only reference the content their role is allowed to see.

Chatbot platforms and RAG products

Offer web ingestion, PDF parsing, and change monitoring as built-in primitives so every customer starts with a KB that stays fresh on its own.

[ 01 / 03 ]
·
Use Cases
AI Assistant
AI Assistant
withFirecrawl
Real-time·Updated 2 min ago

How it works

[ 01 / 06 ]

Crawl your docs, help center, and product sites

Point Crawl at your knowledge sources and Firecrawl returns Markdown or JSON with URLs, headings, and metadata preserved on every chunk, so retrieval starts from the same content your customers see.

[ 02 / 06 ]

Parse the PDFs and whitepapers your team publishes

Send product manuals, whitepapers, contracts, and PDF guides to Parse and Firecrawl returns layout-aware Markdown, so the assistant answers from the docs your team actually publishes, not paraphrased summaries.

[ 03 / 06 ]

Refresh the knowledge base when sources change

Point Monitor at the docs and help pages that back each assistant. Scheduled diffs plus an AI judge fire a webhook when something meaningful changes, so embeddings refresh instead of drifting silently.

[ 04 / 06 ]

Scope content per assistant or workspace

Use Crawl's includePaths and excludePaths controls, then enforce per-workspace allowlists in a thin wrapper so each tenant only pulls from the content it is allowed to use.

[ 05 / 06 ]

Add live web Search for anything outside the KB

Wire Firecrawl Search as a tool call and the assistant answers questions the KB does not cover with a live result plus a citation, instead of an "I don't know" wall.

[ 06 / 06 ]

Show citations in the UI

Every fetch preserves URL, title, and timestamp, so the UI can render a source list and a read-more link per answer without your team writing that plumbing.

[ 02 / 03 ]
·
What Our Customers Say
//
Community
//

People love
building with Firecrawl

Discover why developers choose Firecrawl every day.

How Firecrawl compares to alternatives

FeatureFirecrawlManual CSV uploadsBrowser extensionsGeneric scrapers
Document parsing (PDFs, manuals, whitepapers)YesNoNoNo
Refresh embeddings on source changesYesNoNoNo
Citations with URL and timestamp per chunkYesNoNoNo
Live web search fallbackYesNoNoNo
Structured markdown outputYesNoNoNo
Automatic scheduling & refreshYesNoNoYes
JavaScript renderingYesNoYesNo
URL metadata preservedYesNoNoNo
Multi-tenant scopingYesNoNoNo
API-first integrationYesNoNoYes
Built-in rate limiting & retriesYesNoNoNo
No manual intervention requiredYesNoNoNo
//
FAQ
//

Frequently
asked questions

Everything you need to know about this use case.
General
Firecrawl runs before retrieval, turning your docs, help centers, and product sites into clean Markdown or JSON with URLs attached. Your chatbot or RAG stack uses that output as the source for embeddings, retrieval, and context, instead of raw HTML or manual uploads.
Yes. Upload the file to the Parse endpoint and Firecrawl returns clean, layout-aware Markdown. For a public document URL, use Scrape instead. One call handles PDF, DOCX, DOC, ODT, RTF, XLSX, XLS, and HTML, so product manuals, whitepapers, and contracts feed the assistant the same way as a normal web page.
Technical
Point Monitor at the docs and help pages that back the assistant and set a schedule. Each run diffs against the last snapshot, an AI judge scores the change against a plain-language goal, and a webhook fires only when the change is meaningful, so embeddings refresh from real signal.
Yes. Wire Firecrawl Search into the agent as a tool call and it can query the live web when the KB does not cover a question, then cite the URL it used in the answer, so the assistant does not fall back to a hard no.
Integration
Yes. Scope Firecrawl jobs by domain, path, or custom rules per workspace or customer. Multi-tenant platforms keep every AI assistant grounded only in content it is allowed to reference.
Every Firecrawl fetch preserves the URL, title, and timestamp on every chunk. Your UI can render a source list, a read-more link per answer, and a snapshot per claim without your team writing that plumbing.
Advanced
Constrain retrieval to Firecrawl-collected content with clear URLs, headings, and timestamps, set Monitor to refresh embeddings when the source changes, and surface citations in the UI so users see exactly which page each answer came from.
Documentation portals, help centers, FAQs, status pages, product sites, and any PDF library your team publishes. You choose which domains and paths to crawl based on what each assistant should know.
Why Firecrawl?
The world's most comprehensive web data API. Our custom browser stack and semantic index deliver superior data quality across any website, handling more content types and edge cases than any competitor.
JavaScript rendering, dynamic content, and robust request handling built-in.
Process millions of pages with automatic rate limiting, caching, and distributed infrastructure.
Optimized scraping engine with parallel processing and smart caching for instant results.
Comprehensive docs, SDKs for all major languages, and dedicated support to help you succeed.
[ 03 / 03 ]
·
Pricing

Flexible pricing

Start for free, then scale as you grow.

Free Plan

A lightweight way to get started.
No cost, no card, no hassle.
$0
/month
500 searches or 1,000 pages scraped
2 concurrent requests
Low rate limits

Hobby

Great for side projects and small tools.
Fast, simple, no overkill.
$16
/month
Billed yearly
Save $38
2,500 searches or 5,000 pages scraped
5 concurrent requests
Basic support
$9 per extra 1.5k credits

Standard
Most popular

Perfect for scaling with less effort.
Simple, solid, dependable.
$83
/month
Billed yearly
Save $198
50,000 searches or 100,000 pages scraped
25 concurrent requests
Standard support
$47 per extra 35k credits

Growth

Built for high volume and speed.
Firecrawl at full force.
$333
/month
Billed yearly
Save $798
250,000 searches or 500,000 pages scraped
50 concurrent requests
Priority support
$177 per extra 175k credits

Scale Plans

High-volume plans for teams that need more power and dedicated support. Get access to higher rate limits, more concurrent browsers, and priority support. Scale checks out instantly, no sales call needed.

Need more? Contact us

Scale

For teams scaling their data pipelines
1,000,000 credits / month
$599/month
Billed yearly
Save $1,798
500,000 searches or 1,000,000 pages scraped
100 concurrent requests
Priority support
$397 per extra 350k credits

Enterprise

Power at your pace with custom solutions
Custom credits
Custom concurrent requests
Dedicated support & SLA
Bulk discounts
Zero-data retention
SSO & advanced security