Firecrawl
alternatives.
LLM-ready crawling: crawl or scrape pages straight into Markdown/JSON for LLMs and RAG.
| Option | Model | Starts at | Free | Formats / proxy |
|---|---|---|---|---|
| Credits | $16 / mo | 500 one-time credits | mdhtmljsonscreenshotproxy · tiered | |
| Pay per page | $2 / 1k pages | Free credits on signup | mdhtmljsonproxy | |
| Prepaid tokens | $50 / 1B tokens | 10M tokens with a new API key (non-commercial use only, CC-BY-NC) | mdjsonno proxy | |
| Subscription | $9 / mo | 500 requests (trial, no credit card) | mdhtmljsonscreenshotproxy | |
| Credits | $0 / mo | 500 API credits, one-time (paid plans from $20 / mo) | mdhtmljsonscreenshotno proxy | |
| Pay-as-you-go | $350 / mo | Run requests before you buy (amount not stated on pages visited) | mdscreenshotproxy |
Crawl a whole site and get clean Markdown, HTML or text for LLMs.
Crawl-first design and per-page pricing with no seat or monthly minimum.
Smaller ecosystem than Firecrawl.
Convert any URL to LLM-friendly text by prepending r.jina.ai
Jina Reader is a URL-to-LLM-text endpoint with a free API key tier and a companion search endpoint, which makes it a lightweight option for grounding LLMs.
Billed in tokens, not pages: $50 buys 1B tokens ($0.05 per 1M), and tokens per page vary with content length, so cost per page is not estimated here. The free 10M tokens are non-commercial only (CC-BY-NC). Published rate limits are 20 RPM without a key, 500 RPM with a key and 5000 RPM on premium.
Scrape, crawl, map and search the web via one API for AI agents
Olostep covers scraping, crawling, mapping and search from one API, with monthly plans starting at $9 and a 500-request trial.
The pricing page says plans include successful requests while the FAQ says credits, and the trial's 500 requests are not stated as monthly.
AI scraping API: Markdown, screenshots and JSON extraction via prompts
Useful to compare against providers that sell Markdown or HTML output, because ScrapeGraphAI bills per credit and charges a flat +5 for stealth rather than a multiplier.
Plan 'pages' values are monthly credits, not pages; a scrape costs 1 credit for Markdown, 2 for a screenshot, and extract costs 5 plus any stealth modifier, so credits per page vary by endpoint.
Scrape, crawl and search the web on one API, with its own browser and proxies
Spider combines crawling, scraping, search, a browser API and residential proxies under one key, so it can be compared against both scraping APIs and proxy/browser providers.
Per-endpoint costs, JS and proxy multipliers, page counts per plan and HTML/JSON output support are not stated on the pages visited, so the plans and multipliers are left null until the Scrape and Crawl API docs are checked.