- Who it helps
- Developers building agents that need source-linked public web evidence.
- Availability
- The October 17 announcement links the public documentation. API calls require a DeepSeek API key; no request was made during this review.
What a request returns
Search returns result entries and can request page content. Fetch accepts a page URL. Empty results, missing dates and unavailable content are valid outcomes that an application must handle.
What to budget
The documentation lists $0.75 per 1,000 successful requests for each service, charged when a 200 response contains results or page content. Each account has a five-request-per-second limit per service. Temporary free server-side tool calls in supported harnesses end October 31 at 16:00 UTC; this does not promise free direct API use.
What you can do
- Read the response examples and billing conditions before designing retries, result caching and a request budget. Open reference ↗
Where the evidence stops
Search coverage, extraction quality and latency were not tested. Access restrictions can prevent fetching a page.
Sources read. Public source material was read October 17, 2026 UTC. No model inference or third-party project was executed.
Sources & further reading
Web · official
Search & Fetch API ↗
Public documentation specifies endpoint behavior, request limits and successful-request billing.
DeepSeek · page checked 2026-10-17
Web · official
DeepSeek October 17 release ↗
Dated release identifies the model update, availability, harness requirements and Search & Fetch feature.
DeepSeek · page checked 2026-10-17 · Published 2026-10-17
Coverage note: The cited bundle covers the available primary explanation. Matching original X posts, a complete video comparison and independent execution evidence were unavailable.
The record over time (1)
Dated source checks and published explanations. An observation date can differ from the original release date.
DeepSeek’s new Search & Fetch endpoints let an application retrieve web results and page content. They are separate services with their own request limits and billing rules.