Firecrawl Search vs Kagi Search API for AI agents
Firecrawl Search is a web scraping platform whose search endpoint returns web results and can scrape each result to Markdown in the same call. Kagi Search API is programmatic access to Kagi's premium, ad-free search results for search products and agents.
The short answer
- Firecrawl Search: Search plus full-page scraping to clean Markdown in one request.
- Kagi Search API: High-quality, ad-free results with clear terms that explicitly allow feeding results to an LLM.
Side by side
| Firecrawl Search | Kagi Search API | |
|---|---|---|
| Searches | The web, news and images (the provider does not name the underlying engine). | Kagi's search results (web, news, video, podcasts, images per its MCP). |
| Returns | Title, URL and description per result; optionally full-page Markdown, HTML, links or screenshots for each result. | JSON search results; optionally full Markdown extraction of the top results in the same request. |
| Before the first call | No key to get started (low limits); an account and API key for normal use. | A Kagi account, an API key from the API dashboard, and billing set up (pay-per-use, prepaid credit or invoice). |
| Free | 1,000 credits a month, no card (about 500 searches at 2 credits per 10 results). | Not stated on its pages |
| Paid | Hobby $19/month (or $16/month billed yearly) for 5,000 credits, about 2,500 searches. | $12 per 1,000 requests; Extract $4 per 1,000 pages. |
| Acting on a result | Through the web page each result links to | Through the web page each result links to |
| Limits on keeping or showing results | Not stated on its pages | No storing, caching or building a database of results beyond transient caching; cached results must be deleted on termination. |
| AI in its terms | Not stated on its pages | Sending results to an AI model to answer your users is allowed; training or fine-tuning models, or harvesting training/evaluation datasets, is not. |
| MCP server | Official MCP server: https://mcp.firecrawl.dev/v2/mcp (OAuth: https://mcp.firecrawl.dev/v2/mcp-oauth) | Official MCP server: https://mcp.kagi.com/mcp (code: https://github.com/kagisearch/kagimcp) |
Questions people ask
What do I need before I can use Firecrawl Search?
No key to get started (low limits); an account and API key for normal use.
Is Firecrawl Search free?
1,000 credits a month, no card (about 500 searches at 2 credits per 10 results).
What do I need before I can use Kagi Search API?
A Kagi account, an API key from the API dashboard, and billing set up (pay-per-use, prepaid credit or invoice).
Is Kagi Search API free?
Its pages listed no free use when we read them on October 15, 2026.
Sources
Every line about another company comes from its own pages. Prices and terms change, so check the source before you rely on one.
Firecrawl Search, read on October 15, 2026:
- firecrawl.dev/pricing: “1,000 credits / month”
- firecrawl.dev/pricing: “No cost, no card, no hassle.”
- firecrawl.dev/pricing: “500 searches or 1,000 pages scraped”
- firecrawl.dev/pricing: “Hobby: $19/month billed monthly, or $16/month billed annually, for 5,000 credits per month”
- firecrawl.dev/pricing: “Search costs 2 credits per 10 results.”
- docs.firecrawl.dev/features/search: “No API key needed to get started.”
- docs.firecrawl.dev/features/search: “full-page markdown, HTML, links, or screenshots for each result”
- docs.firecrawl.dev/mcp-server: “https://mcp.firecrawl.dev/v2/mcp-oauth”
Kagi Search API, read on October 15, 2026:
- kagi.com/api/pricing: “$12 /1k requests”
- kagi.com/api/pricing: “$4 /1k pages”
- kagi.com/api/docs: “The Search API is a feature-rich endpoint for search products and agents.”
- kagi.com/api/docs: “You can also request full markdown extraction of the top results in a single API request.”
- help.kagi.com/kagi/api/overview.html: “The API uses a pay-per-use model”
- kagi.com/privacy/api: “Store, cache, or build a database of Results beyond the transient caching needed to operate your application.”
- kagi.com/privacy/api: “including sending Results to an AI model to answer your users' queries.”
- kagi.com/privacy/api: “Use Results to train or fine-tune a search index, ranking system, translation or summarization model”
- github.com/kagisearch/kagimcp: “The Official Model Context Protocol (MCP) server for Kagi Search & other tools.”