| Type | Startup | Startup | Startup |
|---|
| CategoriesDiffers | | | |
|---|
| In one line | Search API purpose-built for AI agents. | Search and extraction API optimized for LLMs and agents. | Turn websites into clean, LLM-ready markdown. |
|---|
| Stage & funding | — | — | — |
|---|
| GitHub stars | — | — | — |
|---|
| License | — | — | — |
|---|
| Language | — | — | — |
|---|
| Last push | — | — | — |
|---|
| Works with | — | — | — |
|---|
| Built on | — | — | — |
|---|
| Built on it | — | — | — |
|---|
| Similar toDiffers | | | |
|---|
| People linkedDiffers | 2 | 2 | 3 |
|---|
| Events | 0 | 0 | 0 |
|---|
| Pros | - +User reportsRuns its own large, independent web index (confirmed on HN by Exa's head of index) rather than reselling Google or Bing.[1]
- +User reportsIn one user's side-by-side test, LLMs preferred Exa and Tavily results over Brave and SearXNG.[1]
- +User reportsBuilt into agent tools: OpenCode supports Exa search by default, and Exa is available as a Codex plugin.[1][2]
- +DocsThe free Starter plan gives $10 of credits a month plus a $10 onboarding bonus, with no payment method needed.[1]
| - +DocsSimple credit pricing: 1,000 free credits a month with no card, and a basic search costs 1 credit.[1]
- +DocsWidely adopted: Nebius cites a developer community of over a million users, and customers including IBM, Cohere and Groq.[1]
- +User reportsUsers rate its results alongside Exa's and like its easy CLI and skill.[1][2][3]
- +DocsOne API covers search, extract, map, crawl and research, and failed URL extractions are never charged.[1]
| - +MetadataVery popular (about 197k GitHub stars) and actively maintained, with a core you can self-host.[1]
- +User reportsTurns pages into clean, LLM-ready Markdown; builders use it in production scraping pipelines and recommend it alongside Exa.[1][2]
- +User reportsSelf-hosting works well for one agent user who does not need the cloud-only features.[1]
- +DocsThe free plan gives 1,000 credits a month with no card, and Hobby is $16/month billed yearly for 5,000 credits.[1]
|
|---|
| Cons | - −DocsPricier than some rivals: search starts at $4 per 1,000 requests (users quoted $7 per 1,000 plus a charge per result beyond the top 10), against $0.25-5 elsewhere.[1][2][3]
- −DocsContents cost $1 per 1,000 pages for each content type, so asking for text, highlights and summaries multiplies the cost.[1]
- −DocsZero data retention and HIPAA are Enterprise-only, which one user described as Exa charging for no retention.[1][2]
- −DocsThroughput is capped at 10 search QPS on the free plan and 25 on Developer; higher QPS and SLAs are Enterprise-only.[1]
| - −DocsNebius, an AI cloud company, agreed to acquire it in February 2026, so roadmap and pricing may follow Nebius's platform priorities.[1]
- −Independent testA competitor's open NEEDLE benchmark puts Tavily below Parallel, Exa and DeepSeek search on quality.[1]
- −DocsResearch requests are priced dynamically at 15-250 credits each on the pro model, and pay-as-you-go costs $0.008 per credit.[1]
- −User reportsIt is a hosted-only API, so 'self-hosted' agent stacks that use it still need an external key.[1]
| - −DocsAGPL-3.0, and the default self-hosted stack has no Fire-engine anti-bot, no screenshots, no page actions and no Agent or Browser features. Its quickstart turns API auth off and has no high availability.[1][2]
- −DocsUnused credits roll over only on Scale and Enterprise, and the JSON, Question and Highlight formats add 4 credits per page.[1]
- −User reportsSite operators complain about its traffic: one said Firecrawl customers caused 95% of their bandwidth overage and that it is not a 'verified bot'.[1]
- −User reportsA competitor says it often struggles with heavy SPAs and hardened anti-bot walls such as Turnstile and DataDome.[1]
- −User reportsA privacy-focused user lists Firecrawl among the providers that keep request logs.[1]
|
|---|
| Best for | Agents that need high-quality semantic web search from an independent index. | Agent builders who want one simple API for search and page extraction that is cheap to start with. | Agents and RAG pipelines that need web pages as clean Markdown at moderate volume. |
|---|
| Watch out for | Model your costs per request and per content type, and get zero retention in writing if your queries are sensitive. | Watch for pricing and roadmap changes after the Nebius deal, and cap how many credits research requests can use. | Expect blocks on protected sites, and check the AGPL terms before you embed it. |
|---|
| Links | | | |
|---|