DeepSeek has added two standalone tools to its API platform: a Search endpoint that queries the web and returns structured results, and a Fetch endpoint that retrieves a single page and returns its content as Markdown. According to the company's Chinese-language API documentation, the tools are meant to give agent and retrieval-augmented generation (RAG) applications access to live web content without a separate third-party search provider.
Both are raw JSON APIs that use the same Bearer authentication as the rest of the DeepSeek API. Search is served at api.deepseek.com/v1/tools/search and Fetch at api.deepseek.com/v1/tools/fetch.
A Search request takes a query of up to 128 characters and can return between 1 and 50 results, with 10 as the default. Developers can pass the user's original question to help the service read intent, supply an IP address to favor localized results, and ask for full page text. Queries accept operators including site: and -site: to include or exclude domains, quoted exact phrases, lang: to restrict languages, and after: and before: to bound publication dates. Each result carries a title, URL, a short snippet of roughly 100 to 200 characters for display, a longer summary of roughly 600 to 800 characters intended for models to read, an optional full-text field, a publication date, a rerank score from 0 to 3 and an authority rating from 0 to 4.
Fetch takes a single HTTP or HTTPS URL and returns the page title plus Markdown body. DeepSeek notes that results depend on the target site's availability and access policies, and that blocked or unextractable pages return an error.
Pricing is per request rather than per token or per result. Search and Fetch each cost 5 yuan per 1,000 successful requests. Only calls that return HTTP 200 with content are billed: a Search must return a non-empty result list and a Fetch must return page text. Failed requests, empty searches and timeouts are free, and the number of results requested does not change the price.
Searches triggered through the Web Search tool in Claude Code or the web_search feature in Codex will not incur Search fees for the first two weeks after the Search API launch. DeepSeek says standard pricing will apply afterwards and that it will announce the exact billing start date separately.
Rate limits are set per account, not per key. Search and Fetch are each limited to 5 queries per second, tracked independently of each other and of model concurrency limits. Fetch also applies per-domain limits, and exceeding either cap returns HTTP 429.
SEE ALSO: OpenCode Makes DeepSeek V4.1 Flash Free to Use · DeepSeek Open-Sources Harness Agent Runtime