DeepSeek Harness v0.2.1-rc.2: V4.1-Flash-1017 and Search/Fetch APIs
The October 17 release candidate adds DeepSeek-V4.1-Flash-1017 support and optional Search and Fetch APIs.
On this page
Release details
· Release candidate
Rc.2 supports DeepSeek-V4.1-Flash-1017. Search returns web search results, and Fetch retrieves a page as Markdown. You can enable both APIs in Web search settings.
Enable the Search and Fetch APIs
- Open Plugins → Built-in plugin settings → Web search.
- Turn on Use the new Web Search API and save.
- Start another search or page fetch. The saved selection applies to that operation.
The switch selects deepseek-search-api and deepseek-fetch-api. The default selection uses Messages search and local HTTP fetch. Turn the switch off and save to restore it.
API base and gateway settings
baseURL defaults to https://api.deepseek.com/v1/tools. Requests append /search or /fetch.
POST https://api.deepseek.com/v1/tools/search
POST https://api.deepseek.com/v1/tools/fetchA custom gateway must route both endpoints. Enter the full base and deployment prefix, such as https://gateway.example/api/v1/tools, without the operation name. Older gateways that only serve model endpoints need these routes added.
The plugin does not automatically retry or switch providers after a failure. Troubleshoot credentials and gateways by error code →
Install and upgrade
CLI users can launch rc.2 with:
npx @deepseek-ai/dsh@0.2.1-rc.2 web