Lightweight fork of DDDMUC/dsh-free-search: multi-engine web search and fetch with automatic fallback, failure cooldown and API key rotation; takes over DSH's default web_search/web_fetch providers. Cannot be installed alongside DDDMUC/dsh-free-search (same package name and bundle row).
Install
# from GitHub (first run asks for allowBuilds approval — follow the hint, retry)
dsh plugin --profile web add github:z-imagine/dsh-free-light-search
Any plugin you install runs third-party code with your own permissions — it can read your files, use your credentials, and reach the network, and tool approvals don’t sandbox it. GitHub-sourced plugins also run build scripts at install time — pnpm blocks those until you allow them, so an install can stop with ERR_PNPM_GIT_DEP_PREPARE_NOT_ALLOWED or ERR_PNPM_IGNORED_BUILDS; dsh prints the exact key to add under allowBuilds in your profile’s pnpm-workspace.yaml, and the install works on the next run. Allowing a build is a trust decision: only install sources you trust, and pin a commit (github:owner/repo#sha).
README
This plugin publishes its README in Chinese only.
轻量纯 AI 搜索插件,面向 DeepSeek Harness (dsh),零成本起步,多引擎回退。 插件注册为 ctx.web seam 的搜索 provider,内置 web_search 工具自动选用,支持引擎切换、回退链、失败冷却与诊断。
中文
设计理念
与抓取型搜索插件不同,本插件专注于纯 AI 搜索 API:调用搜索引擎的结构化 API,返回标题、链接和摘要,不做网页抓取、正文提取或扩展搜索。轻量、快速、低成本。
核心特点
- 纯 AI 搜索:调用搜索引擎 API,返回结构化结果,不做抓取或正文提取
- 免费优先:所有引擎都有免费额度,无 API key 也能用(匿名额度)
- 自动回退:任何引擎失败自动切换下一个,搜索永不中断
- 失败冷却:按失败类别冷却引擎,避免反复撞墙
- 网页设置页:引擎切换、API key 配置、一键测试、引擎开关
引擎列表
本插件只保留有免费额度的 AI 搜索引擎,按额度模型分层:
| 档位 | 引擎 | 说明 |
|---|---|---|
| ① 无限免费 | ddg / ddg-lite / bing / searxng / anysearch |
HTML 抓取型搜索,无限免费 |
| ② 匿名额度 | exa / tavily / firecrawl / you |
无需 key 即可使用,配 key 提升额度 |
| ③ 免费额度 | baidu / jina |
需 key,有免费额度 |
引擎详情
| ID | 引擎 | 免费额度 | 说明 |
|---|---|---|---|
ddg |
DuckDuckGo HTML | 无限 | 偶发限流,自动恢复 |
ddg-lite |
DuckDuckGo Lite | 无限 | 轻量版 |
bing |
Bing | 无限 | 免 key,中文优化 |
searxng |
SearXNG 元搜索 | 无限 | 多实例自动切换 |
anysearch |
AnySearch AI | 无限 | AI 搜索,匿名额度 |
exa |
Exa | 1000 次/月 | AI 搜索,质量高 |
tavily |
Tavily | 1000 次/月 | 专为 AI agent 设计 |
firecrawl |
Firecrawl | 500 页面/月 | 网页抓取 + 搜索 |
you |
You.com | 1000 次/月 | You.com AI 搜索 |
baidu |
百度 AI 搜索 | 每日 50 次 | 中文全网 |
jina |
Jina AI | 有免费额度 | 搜索 + 摘要 |
安装
git clone https://github.com/z-imagine/dsh-free-light-search.git
dsh plugin --profile web add /path/to/dsh-free-light-search
dsh web
使用
网页设置
安装后打开配置页:左侧 插件 页 → 已安装 分组 → free-search → 配置入口。
配置页提供:
- Search engine:下拉框切换引擎
- API keys:为付费引擎填写 key(Exa / Tavily / Firecrawl / You.com / SerpBase / Serply / Baidu / Jina)
- Test engine:测试当前引擎可用性
- 引擎开关:启用/禁用引擎,调整回退优先级
聊天框切换引擎
输入 /free-search-engine,弹出引擎选择窗口,点选即切换。
回退链
默认回退顺序:exa → tavily → firecrawl → serpbase → serply → you → baidu → jina → bing → anysearch → ddg → ddg-lite → searxng
任何引擎失败(限流、反爬、缺 key、网络错误)自动尝试下一个。
失败分类与冷却
quota(额度耗尽):冷却到会话结束auth(key 无效):冷却到会话结束bot-wall(反爬):冷却 60stransient(超时/5xx/限流):重试一次再回退invalid-response(解析失败):不冷却,直接回退
配置文件示例
- id: web
config:
searchProvider: ddg
fetchProvider: fs-quality-fetch
English
Design Philosophy
Unlike scraping-based search plugins, this plugin focuses on pure AI search APIs: calling structured search APIs to return titles, links, and snippets, without web scraping, content extraction, or extended search. Lightweight, fast, and low-cost.
Key Features
- Pure AI Search: Call search engine APIs, return structured results, no scraping
- Free First: All engines have free tiers, usable without API keys (anonymous tiers)
- Auto Fallback: Automatically switch to next engine on failure, search never stops
- Failure Cooling: Cool down engines by failure category, avoid repeated failures
- Web Settings: Engine switching, API key configuration, one-click testing
Engine List
Only AI search engines with free tiers are included, categorized by quota model:
| Tier | Engines | Description |
|---|---|---|
| ① Unlimited Free | ddg / ddg-lite / bing / searxng / anysearch |
HTML scraping search, unlimited free |
| ② Anonymous Quota | exa / tavily / firecrawl / you |
No key required, key increases quota |
| ③ Free Quota | baidu / jina |
Key required, has free quota |
Engine Details
| ID | Engine | Free Tier | Description |
|---|---|---|---|
ddg |
DuckDuckGo HTML | Unlimited | Occasional rate limiting |
ddg-lite |
DuckDuckGo Lite | Unlimited | Lightweight version |
bing |
Bing | Unlimited | No key, Chinese optimized |
searxng |
SearXNG Meta Search | Unlimited | Multi-instance auto-switch |
anysearch |
AnySearch AI | Unlimited | AI search, anonymous quota |
exa |
Exa | 1000 queries/month | High-quality AI search |
tavily |
Tavily | 1000 queries/month | Designed for AI agents |
firecrawl |
Firecrawl | 500 pages/month | Web scraping + search |
you |
You.com | 1000 queries/month | You.com AI search |
baidu |
Baidu AI Search | 50 queries/day | Chinese web search |
jina |
Jina AI | Free tier available | Search + summarization |
Installation
git clone https://github.com/z-imagine/dsh-free-light-search.git
dsh plugin --profile web add /path/to/dsh-free-light-search
dsh web
Usage
Web Settings
Open the config page: Left sidebar Plugins → Installed → free-search → Config entry.
The config page provides:
- Search engine: Dropdown to switch engines
- API keys: Enter keys for paid engines
- Test engine: Test current engine availability
- Engine switches: Enable/disable engines, adjust fallback priority
Chat Command Switching
Type /free-search-engine to open the engine selection popup.
Fallback Chain
Default fallback order: exa → tavily → firecrawl → serpbase → serply → you → baidu → jina → bing → anysearch → ddg → ddg-lite → searxng
Any engine failure (rate limit, bot wall, missing key, network error) automatically tries the next engine.
Failure Classification & Cooling
quota(quota exhausted): Cool down until session endsauth(invalid key): Cool down until session endsbot-wall(anti-bot): Cool down 60stransient(timeout/5xx/rate limit): Retry once then fallbackinvalid-response(parse failure): No cooldown, direct fallback
Config File Example
- id: web
config:
searchProvider: ddg
fetchProvider: fs-quality-fetch
Links
More in this category
DayuanJiang/next-ai-draw-io#dsh-plugin★ 36802
Draws draw.io diagrams in DeepSeek Harness through the next-ai-drawio MCP server: the link in the reply opens a live preview in the right sidebar, you can edit the diagram by hand, and each AI change is a version you can restore. Node.js 18+; npx downloads the MCP server on first use.
Tencent/WeKnora#dsh-weknora★ 36302
Four read-only tools over a WeKnora knowledge base: list knowledge bases, hybrid passage search, reassemble one document's chunks in order, and WeKnora's own cited RAG or ReAct-agent answer with a resumable session id.
superdesigndev/treg★ 7042
Tool catalog for agents: search ~2,600 external endpoints (SEO and SERP, backlinks, social, people and company enrichment, ad libraries, scraping) by the task you want done, read each one's parameters and per-call price, then call it with the credential injected server-side. Ships the skill plus an MCP row that stays disabled until TREG_TOKEN is set.
T-Auto/dsh-ops★ 5074
Provides bash, PowerShell 7, and high-performance Rust tools for DSH on Windows to reduce token usage.
sixtysevenlf/dsh-blender-plugin★ 1575
Lets a DSH agent drive a local Blender over a direct TCP channel with 15 tools: viewport frames, custom-angle renders, Python execution, headless and job offload, transaction rollback, and a planning layer of 228 ops across 34 families.
TencentCloudBase/CloudBase-AI-Toolkit#dsh-plugin★ 1155
Tencent CloudBase backend for DeepSeek Harness — scaffold and deploy full-stack apps from chat, render query results as table cards with paging, sorting and CSV export, preview a deployment on its domain, and call the CloudBase MCP toolset (`mcp__cloudbase__*`) with device-code login.
Community comments
Comments are public GitHub Discussions. Loading them connects to GitHub and Giscus; a GitHub account is required to post.