DeepSeek opens its in-house search engine as Search and Fetch APIs for agents: $0.75 per 1,000 successful requests

Why it matters

【Key points】

On Oct 17 DeepSeek's official API changelog announced it is opening up its in-house search engine as an API. The Search API queries DeepSeek's own engine and returns structured results; the Fetch API retrieves a single page as Markdown. Both target agents and RAG apps and use standard DeepSeek API key auth. DeepSeek says the engine already powers search on its web app and API, handles hundreds of millions of requests a day, indexes hundreds of billions of pages, and filters SEO spam, GEO manipulation and low-quality AIGC.

Both endpoints cost $0.75 per 1,000 successful requests; failures, empty results and timeouts are not billed, with a 5 QPS limit per account for each. Server-side tool calls made when using the DeepSeek API through DeepSeek Harness, Claude Code or Codex already hit these APIs; they are free for now and become billable from 00:00 Beijing time on Nov 1, 2026.

【Verifiable】

Official changelog: https://api-docs.deepseek.com/updates

Search & Fetch guide: https://api-docs.deepseek.com/guides/search/

【Our read】

1. Few model labs sell their own search engine as a standalone API; DeepSeek is stepping straight into agent retrieval and competing with existing search-API vendors.

2. $0.75 per 1,000 calls is cheap for the category, and paired with the same-day V4.1-Flash-1017 upgrade it is attractive for developers building agents on DeepSeek.

3. The index size and the spam and GEO filtering are DeepSeek's own claims; result quality and coverage still need independent testing.