Sunday, October 18, 2026
Herald of AI

AIGenerated AI news, written by AI and explained simply. Models, technology, culture and the people behind it.

How it works

DeepSeek launches Search and Fetch APIs for $0.75 per 1000 requests

The new tools let AI agents pull live web content directly through code, without building their own web crawlers.

AIGenerated · Written by AI from the sources listed below
DeepSeek's new Search and Fetch APIs let AI apps access live web content through simple API calls.
DeepSeek's new Search and Fetch APIs let AI apps access live web content through simple API calls. Picture: api-docs.deepseek.com

DeepSeek has added two new API tools that give AI applications direct access to live web content: a Search API and a Fetch API.

The two tools are aimed at AI agents and RAG systems, which are programs that use retrieved information to improve their answers. Both return raw JSON data and use the same bearer authentication as the rest of the DeepSeek API 1.

Search API

The Search API takes a text query and returns structured results from DeepSeek's in-house search engine. By default it returns up to 10 results, but developers can ask for between 1 and 50. The optional need_content flag lets callers also receive the full page text, though by default only titles, summaries and URLs come back.

Searches support common query operators such as restricting results to a specific site, excluding certain terms or phrases, limiting results by language, and filtering by publication date. Each result includes a rerank score from 0 to 3, with scores of 1.5 or above described as high quality, and an authority tier from 0 to 4 where 3 and above marks highly authoritative sources 1.

Fetch API

The Fetch API takes a single URL and returns the page content converted into Markdown. It accepts HTTP and HTTPS links up to 8192 characters and does not support URLs with usernames or passwords. The response includes the page title and the Markdown content 1.

Pricing

Both APIs are priced at $0.75 per 1000 successful requests. The endpoint for search is POST to /v1/tools/search and for fetch it is POST to /v1/tools/fetch, both under the api.deepseek.com base URL 1.

DeepSeek has been expanding its developer ecosystem beyond its language and reasoning models, offering tools that make it easier for developers to build applications that rely on real-time information.

The documentation was published on the DeepSeek API docs site and includes examples in both Python and curl.

Key facts

  • Search API returns structured web results from DeepSeek's in-house search engine
  • Fetch API converts a single webpage into Markdown
  • Pricing is $0.75 per 1000 successful requests
  • Both are designed for AI agents and RAG applications
  • Query operators support site filters, date ranges, language codes and phrase matching

Sources

  1. Search & Fetch | DeepSeek API Docs
  2. Web scraping — en.wikipedia.org