Web Data API Free Plan | Decodo
This is a sales page for Decodo's Web Data API, which offers a free starter tier with $5 in credit and up to 10K requests a month across four endpoints (/scrape, /search, /map, /crawl). It pitches the API as a single-key tool for pulling clean, LLM-ready web data for AI agents and pipelines.
- The free plan gives $5 credit and up to 10K requests per month with no card required, covering all four endpoints.
- Output comes as Markdown or JSON with nav bars and ads stripped, aimed at RAG pipelines and LLM context windows.
- It ships an official MCP Server plus LangChain and LlamaIndex integrations, and claims a 99.99% success rate with pricing from $0.15 per 1K requests.
Decodo is pitching its Web Data API as a free-to-start scraping service. The starter plan gives you $5 in credit, up to 10K requests a month across four endpoints: /scrape for individual pages, /map for discovering a site's URLs, /crawl for pulling a whole site, and /search for live web queries. No card is required, and the page claims pricing stays flat, with no multipliers for JavaScript rendering or protected sites. That last point matters if you've been burned by scraping vendors whose advertised per-request price balloons once you hit a Cloudflare-protected target. Paid usage starts at $0.15 per 1K requests, and you're only billed for successful responses. The site also cites a 99.99% success rate, 200 requests per second, and a 125M+ IP pool, though those are vendor figures you'd want to check against your own targets.
The more concrete selling point is the output and the agent tooling. Responses can come back as HTML, JSON, CSV, or Markdown, and the Markdown output strips nav bars, cookie banners, and ads, which is the part that saves real cleanup work if you're feeding a RAG pipeline or an LLM context window. Decodo ships an official MCP Server and CLI, and says it's listed in GitHub's official MCP Registry. LangChain and LlamaIndex integrations are included, along with connectors for n8n and OpenClaw, so an agent can call the API as a tool without custom wrapper code. There are also 100+ pre-built templates. The comparison table positions this against "other scraping APIs," claiming those are mostly built around single-page scraping and need custom setup for MCP. That's a self-serving framing, so treat it as marketing rather than benchmarked fact.
The practical takeaway: if you're prototyping an AI agent or a data pipeline and want one key covering discovery, crawling, and extraction, the free tier is a low-risk way to find out whether the output quality holds for your specific sites. The 10K-request cap is enough to map and sample a real site end to end, but it's not enough for sustained production volume, and the page gives no independent verification of its success-rate or anti-bot claims. Test against the sites you actually care about before committing.