Back to vendors
F

Firecrawl

Visit site
Entry priceFrom $16/mo billed yearly · free tier + open sourceFull pricing detail

Web data API for AI agents that searches, scrapes, crawls and acts on web pages through hosted browsers, with an open source core.

Firecrawl is a web data API for AI agents. It searches the web, scrapes pages into clean markdown or structured JSON, crawls whole sites, maps a site's URLs, parses PDFs and Office documents, and lets an agent act inside a page through a hosted browser. The output is shaped for a model's context, stripping navigation and boilerplate.

The endpoints cover the path from finding to using web data. Search returns full page content, not just links. Scrape turns any URL into markdown or JSON, with change tracking between scrapes. Crawl and Map traverse a site. Interact continues on a scraped page by prompt or Playwright code, and Browser Sandbox opens standalone sessions where an agent fills forms, clicks and signs in, with persistent profiles that keep a login between runs. Monitor checks pages, sites or searches on a schedule and notifies a webhook, email or Slack when something changes. The /agent endpoint, a research preview, gathers data from a prompt without known URLs.

Since September 2026 Firecrawl also runs Alexandria, a library of more than 300 data sources that an agent reaches through the same API and through the Firecrawl plugin in Claude, Claude Code, ChatGPT, Codex and Cursor.

It holds partner providers such as Apollo, FullEnrich, Shopify, Benzinga and Fiscal.ai beside Firecrawl's own indexes: a Developer Index of more than 70 million repositories, docs, issues and pull requests, a Research Index of papers that includes 41 million in the life sciences, and a Government and Legal Index of 30 million US government pages. Universal Scrape reaches web pages, files and Alexandria providers through one call. Firecrawl announced Alexandria with a $75 million Series B, and it is also sold through AWS Marketplace and Google Cloud Marketplace.

Firecrawl ships SDKs in nine languages, a CLI, an MCP server and skills for coding agents. The core is open source under the AGPL 3.0 license and can be self hosted with Docker Compose; Agent, Browser, Interact, the dashboard and enterprise controls stay on the cloud service. Enterprise adds SOC 2 Type II documentation, SSO and SCIM, IP and key restrictions, Threat Protection policies on every fetched URL, SIEM audit logging and Zero Data Retention.

Pricing is a credit based subscription. A free tier gives 1,000 credits a month, then Hobby is sixteen dollars a month billed yearly (nineteen monthly), Standard eighty three, Growth three hundred thirty three and Scale five hundred ninety nine, with Enterprise custom. A basic scrape costs one credit per page, while structured formats, search, browser minutes and monitor checks cost more, and error pages still cost a credit.

From our comparison video

Exa vs Firecrawl: One Finds the Pages, One Turns Them Into Data

Our 4 minute video puts Firecrawl side by side with Exa. It covers what each one does and how each is priced, then when to choose which.

Recorded in October 2026. The profile here is live and updates as research is verified.

Vendor details

Canonical URL

https://www.firecrawl.dev

Category

Agent infrastructure

Subcategory

Web scraping and crawling

Funding status

Independent. The core is open source under the AGPL 3.0 license (github.com/firecrawl/firecrawl), offered as a managed cloud service and self hostable with Docker. SOC 2 Type II per its enterprise documentation.

Company status

independent

Use cases & customers

Primary use cases

web scrapingweb crawlingstructured data extractionRAG ingestionweb data for agents

Target customers

developersenterprise

Deployment options

SaaSself-hosted

Integrations

REST API with OpenAPI specifications, SDKs for Python, Node, Go, Java, Ruby, Rust, .NET, PHP and Elixir, a CLI, an MCP server (keyless, OAuth or API key) and skills for Claude Code and Cursor. Listed integrations include LangChain, LlamaIndex, Mastra, the Vercel AI SDK, n8n, Make, Zapier and Stripe Projects; SIEM audit events stream to Microsoft Sentinel.

In practice

Your cold outbound needs a personalized first line per prospect. You scrape each company's homepage and about page with Firecrawl, feed the clean markdown to an LLM, and generate specific opening lines at scale.

Your RAG knowledge base is fed by brittle scrapers that break on JavaScript heavy sites. You replace them with Firecrawl, which renders the page and returns clean markdown ready for ingestion.

You need to know when a competitor changes its pricing page. You set a Firecrawl Monitor on the page with a daily schedule and get a webhook or Slack alert only when the page changes.

Agentic Index coverage score

8.5 / 14 capabilities · 61%

Integrations & Tool Calling Partial

An agent gets the web as tools through the API, SDKs, CLI and an MCP server. The tools search, scrape, crawl, map, parse and monitor, and include Firecrawl's research agent, and Firecrawl is listed with LangChain, LlamaIndex, n8n, Make and Zapier. It connects to websites, not to the customer's other systems. There are no native connectors that act in a CRM, ticketing or messaging tool, and no per user credentials to scope, rotate or revoke for them. Acting inside a web page happens through the hosted browser, not through connectors.

Sourcefirecrawl.dev/llms.txtread 2026-09-22

Workflow Orchestration Not documented

Customers get no workflow runtime of their own. The /agent endpoint searches, navigates and gathers data by itself from a prompt, optional URLs and a schema, but that pipeline is Firecrawl's own prebuilt agent, and the customer cannot define, branch or sequence its steps. Crawl, batch scrape and monitors are jobs Firecrawl runs to a fixed pattern. Nothing lets the customer sequence, branch, retry or route work while mixing deterministic nodes with agent steps. The customer's own logic lives in its framework or in tools such as n8n.

Sourcedocs.firecrawl.dev/features/agentread 2026-09-22

Knowledge Grounding & RAG Not documented

The public web is what Firecrawl searches, scrapes and crawls, and its Research and Developer indexes cover public papers and developer artifacts. None of it is a structure Firecrawl maintains over the customer's own documents. Parse turns a customer's PDFs and Office files into markdown and JSON for the customer to use, which is extraction, not an index that persists and stays queryable. Firecrawl does not index, refresh or permission customer content. Clean output feeding the customer's own RAG layer grounds that layer, not Firecrawl.

Sourcefirecrawl.dev/llms.txtread 2026-09-22

Human Oversight & Guardrails Partial

Threat Protection lets an organization set one policy that every URL is checked against before Firecrawl fetches it, whether a scrape target, a search result, a link found during a crawl, or an agent's starting URL. The policy takes a custom blocklist and allowlist of domains, blocked top level domains, a risk score threshold and a setting to fail closed.

It can be driven by the organization's own Zscaler categories and locked so no request weakens it. The check is enforced before the fetch, but there is no approval step, checkpoint or escalation to a person. Threat Protection is an enterprise feature enabled per organization.

Sourcedocs.firecrawl.dev/features/threat-protectionread 2026-09-22

Security, Identity & Governance Full

The SOC 2 Type II report is independently audited, and the security page carries the SOC 2 Type 2 mark. Controls for customers sit beside it, with SAML and OIDC single sign on through Okta, Entra ID or Google Workspace, SCIM provisioning and deprovisioning, and team roles. API keys can be restricted to IP ranges or locked to specific endpoints and formats, Zero Data Retention covers scrape and search, PII redaction is available, and DPAs and custom contracts are offered. The enterprise controls are provisioned per organization.

Sourcedocs.firecrawl.dev/enterpriseread 2026-09-22

Observability & Auditability Full

Each step the jobs and agent take is recorded. SIEM Audit Logging streams one structured event for every URL Firecrawl fetches for the customer, whether from a scrape, crawl, batch, search, extract, parse or agent run.

Each event carries the target, status and outcome, including fetches blocked by policy, the API key and workflow that caused it, a request identifier that groups a whole crawl or agent run, and the customer's correlation IDs, and the events go to Microsoft Sentinel. The dashboard's Activity Logs show request history with status, duration and credits. The log is kept apart from the runtime and exported to the SIEM, so every tool call has a record. Page content is not logged, and Firecrawl does not say how long logs are kept on each plan. SIEM logging is an enterprise feature enabled per organization.

Sourcedocs.firecrawl.dev/features/siemread 2026-09-22

Memory & State Persistence Not documented

The agent has no memory it reads to decide. The closest thing is the saved login. A scrape can save its browser session to a named profile that later scrapes and Interact sessions load, and Browser Sandbox sessions can persist.

The browser loads cookies, local storage and an authenticated profile when a session starts, so the product applies them as settings and the agent does not read them as context. Nothing keeps context across a run, conversation or workflow, there is no memory layer for the longer term, and nothing lets memories be reviewed, edited, deleted or scoped. Change tracking's comparison with the previous scrape is likewise applied as a rule.

Sourcedocs.firecrawl.dev/features/interactread 2026-09-22

Deployment & Data Residency Full

Customers can use the managed cloud service or run Firecrawl on their own infrastructure. The core is open source under the AGPL 3.0 license and ships a pinned Docker Compose guide for self hosting. A self hosted install exposes the core scraping APIs, with the customer owning authentication, TLS, persistence, monitoring and upgrades and connecting its own provider compatible with OpenAI, or Ollama, for LLM formats. Agent, Browser, Interact, the dashboard and enterprise controls run only in the cloud. The cloud service offers no choice of region.

Sourcedocs.firecrawl.dev/contributing/open-source-or-cloudread 2026-09-22

Prebuilt Agents, Templates & Packs Partial

A single prebuilt agent ships. The /agent endpoint searches, navigates and gathers structured data from a prompt and optional schema, and it is also offered in the dashboard with no code. Firecrawl labels it a Research Preview in early access, and any account can call it. Firecrawl also publishes skills for Claude Code, Cursor and other coding agents covering Search, Scrape, Crawl and Interact, all made for one job, gathering web data. Beyond that one agent in preview and the skill set, there is no catalog of workflows, templates or agents for specific roles to browse and adopt.

Sourcedocs.firecrawl.dev/features/agentread 2026-09-22

Triggers & Channel Coverage Full

Monitor runs recurring checks on a schedule the customer sets, against named URLs, a scheduled crawl of a whole site, or a web search, with an optional goal in plain language. Each check labels every page same, new, changed, removed or error. Notifications go to a webhook per page, a webhook per completed check, an email summary or a Slack channel. Work starts on a schedule with no person initiating it. Crawl, batch and agent jobs also post lifecycle events to the customer's webhook.

Sourcefirecrawl.dev/llms.txtread 2026-09-22

Model Flexibility & Routing Full

Web data goes to whatever model the customer's agent runs, through LangChain, LlamaIndex, Mastra, the Vercel AI SDK and any MCP client, and a self hosted deployment connects its own provider compatible with OpenAI, or Ollama, for formats that use an LLM. On the cloud service the extraction and /agent models are Firecrawl's own, with a choice among its agent models. Model choice sits with the customer for its own agent, and for extraction when self hosted.

Sourcedocs.firecrawl.dev/contributing/open-source-or-cloudread 2026-09-22

APIs, SDKs & MCP Extensibility Full

A REST API makes Firecrawl callable from outside, with published OpenAPI specifications for v2 and for webhooks. Official SDKs cover Python, Node, Go, Java, Ruby, Rust, .NET, PHP and Elixir, and there is a CLI and an MCP server that works keyless, with OAuth sign in or with an API key, plus skills for coding agents.

Sourcedocs.firecrawl.dev/llms.txtread 2026-09-22

Testing, Debugging & Optimization Not documented

The Playground lets a developer try a request and inspect the raw response, and the Debug issue button in Activity Logs runs Firecrawl's support agent on a failed job to suggest a fix. Both test and debug Firecrawl calls, not the customer's agent. There are no fixtures, test datasets, scoring or quality gates for the customer's agent workflows.

Sourcefirecrawl.dev/llms.txtread 2026-09-22

Browser & Computer Use Full

Agents control real browsers hosted by Firecrawl. Browser Sandbox gives a secure, isolated session where agents fill out forms, click buttons and authenticate, with Playwright and agent-browser preinstalled, a CDP URL and a live view. Interact continues on any scraped page by prompt or by Playwright code for flows with several steps, such as signing in and clicking through pagination. The browsers are managed and sandboxed remotely. Browser Sandbox and Interact run only in the cloud.

Sourcedocs.firecrawl.dev/features/browserread 2026-09-22

The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded

Product updates

·MCP / tool calling / APIVerified

The Government & Legal Index is a free search index of 30 million pages from US federal, state and local government sites, covering statutes, regulations, city codes, permits, court opinions and agency guidance. Agents reach it through a new government search endpoint in the API, a matching tool in the Firecrawl MCP server, the CLI, or the Firecrawl plugin in ChatGPT and Claude. Searches use no credits.

Bears on: MCP / tool calling / API

View source
·MCP / tool calling / APIVerified

Universal Scrape extends Firecrawl's scrape endpoint so one call can read web pages and files such as PDFs, images, Word documents, spreadsheets and presentations, or call an Alexandria data provider such as Apollo, FullEnrich, Data Legion, Fiscal.ai, Benzinga or Particle. Calls use the same Firecrawl key and credits, and some providers need an org admin to accept their terms before the first call.

Bears on: Integrations

View source
·MCP / tool calling / APIVerified

Firecrawl launched Alexandria, a common interface for agents to discover and retrieve information from official data providers, custom connectors and its Research, Developer and Government indexes. Agents can query providers and use specialized tools alongside live web search through Firecrawl.

Bears on: MCP / tool calling / API

View source
View all 5 Firecrawl updates →Tracked since Jul 2026 · Verified from public vendor sources

Pricing

From $16/mo billed yearly · free tier + open source

Credits per API request (1 credit per page for scrape, crawl, map, monitor), within monthly subscription tiers

Free tier

Included quota

Free 1,000 credits a month. Hobby 5,000 credits, Standard 100,000, Growth 500,000, Scale 1,000,000. One credit is one page on a basic scrape, crawl or map; credits are shared across all endpoints.

What is public

Firecrawl publishes six tiers from Free to Enterprise, monthly and annual prices, and per endpoint credit costs.

Billing mechanics

A credit based monthly or annual subscription; the annual figure is the effective monthly rate, not a separate plan. Each tier grants a monthly credit pool shared across all endpoints, with most calls one credit per page and higher rates for structured formats, search, browser minutes and monitor checks. Annual Scale and Enterprise plans allow some rollover.

Cost watchouts

A page that returns an error status such as 403 or 404 still costs a credit. Structured formats (JSON, Question, Highlight) add 4 credits per page, Search costs 2 credits per 10 results, Interact 2 credits per browser minute, and Monitor 1 credit per page per check.

Variable cost rationale

Tiers are fixed monthly subscriptions with predictable credit pools, but structured formats, search, browser minutes and monitor checks consume credits faster than the one credit per page base, and unused credits are generally lost.

Additional watchouts

Unused credits generally do not roll over. Structured output adds four credits per page, so an extraction-heavy crawl costs several times the one credit base, and error pages still cost a credit.

Overage / add-ons

Credit based subscription tiers; exceed a tier by upgrading or buying more credits. Annual Scale and Enterprise plans allow some credit rollover; other credits do not carry over.

Sales call required

No, self serve available

Free / trial

Free tier: 1,000 credits per month, no card. Open source core self hostable.

Lowest paid plan

Hobby $16/mo billed yearly or $19/mo billed monthly (5,000 credits)

Commercial notes

Sold bottom up to developers and AI teams through a free tier, open source distribution under AGPL 3.0, an MCP server, skills and framework integrations. Self serve tiers run up to Scale; Enterprise adds custom credits, SSO and SCIM, IP and key restrictions, Threat Protection, SIEM audit logging, Zero Data Retention, pooled credits, spend limits and dedicated support.

Key ambiguities

Real monthly cost depends on which endpoints and formats a workload uses, since structured formats, search, browser minutes and monitor checks draw credits at different rates from the one credit per page base.

Cancellation / refund

Self serve monthly or yearly subscriptions via Stripe; standard subscription cancellation. Enterprise terms are contractual.

Support SLA / resale

Self serve support by tier; Enterprise adds dedicated support and SLAs.

Missing data

Enterprise credit volumes and discounts are custom, and Agent run pricing is not on the pricing summary read.

Agentic Index verified 2026-09-22

Alternatives to Firecrawl

The closest documented capability profiles to Firecrawl among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.

  • Bright Data8.0 / 14Fuller documented coverage on Human Oversight & Guardrails
  • Daytona8.0 / 14Adds documented Memory & State Persistence
  • Confident AI8.5 / 14Adds documented Testing, Debugging & Optimization
  • HoneyHive8.5 / 14Adds documented Testing, Debugging & Optimization
  • Langfuse8.5 / 14Adds documented Testing, Debugging & Optimization
  • Arize AI9.0 / 14Adds documented Testing, Debugging & Optimization

Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.