dsh-web-search-free: A Multi-Engine Free Web Search Plugin for DeepSeek Harness
mochinek0/dsh-web-search-free
A multi-engine free web search and fetch provider for DeepSeek Harness, with automatic fallback across Tavily, Firecrawl, Exa, etc., without consuming model tokens.
dsh-web-search-free is a free web search and scraping plugin built for DeepSeek Harness, replacing the official search channel with a multi-engine auto-fallback pipeline to avoid the high cost of burning a full model call per search. Once installed, you simply fill in API keys for each engine in the settings card and reorder them as needed; search and fetch requests are tried in order, seamlessly falling back to the next available engine on failure or quota exhaustion. It requires no LLM credentials and reverts to the default channel on uninstall, making it a practical choice for DSH users seeking low-cost, high-availability search.
How to Install
dsh plugin --profile web add dsh-web-search-free- Category
- Tools & Capabilities
- Platform
- DSH Plugin
- Author
- mochinek0
- Distribution
- Plugin
dsh-web-search-free Key Features
dsh-web-search-free Repo Summary
What Does It Do?
dsh-web-search-free is a DSH plugin for DeepSeek Harness that provides free web search and scraping by replacing the default official search channel with a multi-engine auto-fallback pipeline. It solves the high cost of the official channel, which burns a full model call per search, by calling each search engine's dedicated retrieval endpoints directly without any LLM involvement. The plugin registers as a web capability channel in DSH, offering both searchProvider and fetchProvider, and ships with a draggable settings card that supports Chinese and English interfaces.
Core Features
- Multi-engine auto-fallback: supports TinyFish, AnySearch, Exa, Tavily, Firecrawl, Brave, SerpApi, Jina, and tries them in your configured order, falling back automatically on failure or quota exhaustion.
- Per-engine multi-key rotation: you can enter multiple API keys for the same engine, and they rotate in order to improve reliability.
- Zero model token consumption: all retrieval requests are sent directly from the DSH host process to the engines, bypassing any LLM and saving DeepSeek model tokens.
- Toggleable web_fetch tool: mount or unmount the model-side web_fetch tool directly from the settings card, avoiding useless errors.
- Clean uninstall: installed as a bundle layer, it reverts to the default channel after uninstall and restart, with no manual profile edits.
How to Use This Plugin?
After enabling the plugin in DSH, go to Settings → Plugins → Web Search Free (or 免费 Web 搜索 in Chinese), fill in API keys for the engines you want to use, and adjust their order as needed. The plugin will automatically handle search and fetch requests according to your configured engine order, falling back to the next available engine if one fails or runs out of quota. You can also toggle the web_fetch tool in the card to control whether the model uses web scraping.
This page is an independent rewrite of the plugin's official README — for authoritative documentation and the latest changes, refer to the source: mochinek0/dsh-web-search-free. The plugin is third-party code that runs on your machine once installed; inclusion does not imply endorsement — please review the source before installing.
DSH Plugin FAQ
