Comparison
SerpKite vs Firecrawl
Both turn the web into Markdown for agents. Firecrawl is a scraping platform with search on top; SerpKite is a Google search API with page reading, extract, map and crawl on top.
Updated
We make SerpKite
Summary
The short version
Firecrawl scrapes, crawls and maps sites, fills JSON schemas with an LLM and drives a browser, all from one credit balance on which a search of 10 results costs 2 credits. SerpKite returns Google's actual results page (People Also Ask, knowledge graph, local, news, shopping) at 1 credit per 10 results, reads the top pages as Markdown with include_content, covers sites with /v1/extract, /v1/map and /v1/crawl, and its credits never expire.
Choose Firecrawl if…
- Pages need interaction before they show content: Firecrawl's actions click, scroll, type and wait. SerpKite renders JavaScript pages in a browser but doesn't click or scroll.
- You want screenshots, or DOCX and other office documents parsed. SerpKite reads HTML and PDF and takes no screenshots.
- You want to self-host: Firecrawl is open source under AGPL-3.0.
- You want an agent that finds the data without URLs: Firecrawl's Agent endpoint searches and navigates for you.
Choose SerpKite if…
- Search is the main job, and you want Google's ranking, People Also Ask, the knowledge graph and Google verticals (News, Maps, Shopping, Scholar), not a generic list of links.
- It costs less per search: $0.80 per 1k at entry and $0.60 at best, vs $7.60 on Firecrawl's Hobby plan and $1.20 at its best list price (Scale, billed yearly).
- You want credits that never expire and no subscription. Firecrawl's plan credits reset monthly; only Scale rolls them over, for one month.
-
You don't want to pay for fetches that fail: failed URLs in
/v1/extractand failed or empty searches cost nothing, and short pages that turn out to be bot checks, login walls or soft 404s aren't billed. Firecrawl charges 1 credit for a page that answers 403 or 404. -
You want JSON from pages without paying for failures:
output_schemaon/v1/extractcosts 5 credits per page (1 + 4, the same credit count as Firecrawl's 1 + 4), an extraction that fails costs only the page, and a page that fails costs nothing. -
You want search that keeps answering when Google blocks:
engine=autofalls back to Brave, Bing, DuckDuckGo, Yahoo or Yandex, andengine=consensusmerges several indexes.
Differences
What actually differs
Search source
SerpKite parses Google's logged-out results page by default, and meta.engine names the engine that answered. Firecrawl's docs describe an "upstream search provider" without naming it; results come back as web, news and images lists.
Billing unit
Firecrawl bills one credit balance for everything: 2 credits per 10 search results, 1 per page scraped or crawled, 1 per map call, and +4 per page for JSON extraction. SerpKite bills 1 credit per 10 Google results, +1 per page read with include_content, and 1 per page returned by /v1/webpage, /v1/extract and /v1/crawl.
Search plus page content
Firecrawl's search takes scrapeOptions to return each result's page. SerpKite's include_content returns the top 1–5 pages as Markdown, and highlights keeps only the 3 passages that best match the query, at no extra credit.
Structured extraction
Both fill a JSON Schema from a page with an LLM. Firecrawl's json format adds 4 credits per page; SerpKite's output_schema adds 4, with an optional prompt, written by Claude Haiku 5.5. Schemas of only booleans and enums go to OpenAI's Decisions API, which always returns schema-valid values. summary: true adds a page summary for +4.
Scraping depth
Firecrawl has more browser tooling: actions (click, scroll, type), screenshots and office-document parsing. SerpKite's webpage to Markdown API and extract API (up to 20 URLs or PDFs per call) render JavaScript pages when needed and return clean Markdown, highlights, summaries or JSON, without page interaction.
Crawl and map
Both crawl and map sites. SerpKite's crawler is async, reads up to 1,000 pages per job, respects robots.txt and Crawl-delay, and sends a signed webhook when it ends; map lists a site's URLs from robots.txt, sitemaps and links. Both bill a map as 1 credit per call; SerpKite's is free when it finds no URLs.
Agent tooling
Both ship SDKs and an official MCP server. SerpKite's MCP server is remote (a URL and an API key in Claude, Cursor or ChatGPT), and every result can come back as Markdown or compact JSON to save tokens.
Features
Feature comparison
| SerpKite | Firecrawl | |
|---|---|---|
| Real Google results | Unnamed upstream | |
| Google web + core verticals News, Images, Maps, Shopping, Scholar… | Web, news, images | |
| ChatGPT / Perplexity / Gemini answers SerpKite: ChatGPT, Gemini, Copilot Search, Google AI Mode and Brave | ChatGPT, Gemini (no Perplexity) | |
| Bing / DDG / Brave / Yandex | Bing, DDG, Brave, Yandex | |
| Markdown output | ||
| Resolved final URLs (post-/goto) | Guaranteed | n/a |
| Page content in the same call | include_content | scrapeOptions |
| Async batch + webhooks | 50% off | Crawl, batch scrape |
| Official SDKs TypeScript, Python, Go; LangChain and CrewAI tools | TS, Python, Go | Python, JS and more |
| Official MCP server | ||
| Credits never expire | Monthly plans | |
| Spend caps and per-key limits | Monthly top-up limit | |
| Failed / empty results free | Incl. empty-200 detection | Empty free; 403/404 billed |
| Public status with latency | p50 / p95 | |
| No query logs / DPA SerpKite: No query logs · results cached ≤ 24 h, no identity | Default | ZDR on Enterprise |
| Legal shield | No (honest policy) | |
| SOC 2 | Planned | Third-party audit |
Check = yes, cross = no, dash = not verified. SerpKite rows marked "Planned" are on the roadmap, not live.
Price
Price per 1,000 Google searches
| SerpKite | Firecrawl | |
|---|---|---|
| Free tier | 2,500 on signup + 1,000/mo, no card | 1,000 credits/mo (500 searches), no card |
| Entry, ~$50 pack $ per 1k searches | $0.80 | $7.60 |
| ~100k searches / month $ per 1k | $0.60 | $1.60 |
| ~1M searches / month $ per 1k | $0.60 | enterprise |
| Best public list price $ per 1k | $0.60 | $1.20 |
| Credit expiry | Never | Monthly (Scale: 1-month rollover) |
Public list prices for Google web search, verified on vendor pages on 2026-10-25. "Entry" is the ~$50 pack or plan (SerpKite also has a $10 pack at $1.00 per 1k). SerpKite's self-serve packs stop at 500K credits; larger volumes are quoted on request. Expiring credits raise the effective price if you don't use them in time. See SerpKite pricing.
Firecrawl sells credits, not searches. Its prices here are $ per 1k searches of 10 results (2 credits each): Hobby ($19 a month for 2,500 searches), Growth ($399 for 250,000) and, as the best list price, Scale billed yearly ($599 a month for 500,000). Above Scale it tops up in $5 increments or quotes Enterprise. Scraped pages cost 1 credit each on Firecrawl; on SerpKite, /v1/webpage is 1 credit per page and include_content adds 1 per page to a search.
Verdict
Our take
If your agent has to click through pages, take screenshots or parse office documents, Firecrawl is the more complete scraper, and SerpKite doesn't try to match it. If it mostly searches, reads pages and pulls JSON out of them, SerpKite gives it Google's real results for less per query, Markdown or schema-shaped JSON from the pages, and credits that don't expire. Teams that need both keep Firecrawl for the hard pages and send search and extraction to SerpKite.
Try it
One request to compare for yourself
curl https://api.serpkite.com/v1/search \
-H "Authorization: Bearer $SERPKITE_API_KEY" \
-H "Content-Type: application/json" \
-d '{"q":"best espresso machine 2026","country":"us","format":"markdown"}'1 credit per page of 10 results, SERP features included. Run it in the playground without signing up.
FAQ
Frequently asked questions
Start building
Get your API key in 30 seconds
2,500 free credits, then 1,000 every month. No credit card.