Fetchium v1.0.0 — Public Beta
The universal retrieval layer for humans and AI agents — federated multi-backend search, adaptive content extraction, multi-signal ranking, and cited research, as a CLI / REST API / MCP server.
Highlights
- 🔍 15+ search backends — SearXNG, Google, DuckDuckGo, Bing, Brave, Tavily, Serper, Exa, Firecrawl, ArXiv, Wikipedia, HackerNews, Reddit, GitHub, StackOverflow
- 🤖 AI synthesis — Gemini Flash-powered research pipeline with token-efficient output
- 📡 REST API —
POST /v1/search,/v1/fetch,/v1/scrape,/v1/research - 🔌 MCP server — 12 tools, works with Claude Desktop out of the box
- ⚡ HyperFusion ranking — 8-signal: BM25 + semantic + temporal + authority + evidence + diversity + depth + consensus
- 🌐 Residential proxy — DataImpulse-backed for Google/DDG, automatic IP rotation on block
Install
# npm / npx
npm install -g fetchium-cli
npx fetchium --help
# cargo (from source)
cargo install --git https://github.com/zuhabul/Fetchium fetchium-cli
# cargo binstall (pre-built binary)
cargo binstall fetchium-cliAPI
curl https://api.fetchium.com/v1/search \
-H "Authorization: Bearer YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{"query":"rust async programming","tier":"detailed"}'MCP (Claude Desktop)
{"mcpServers": {"fetchium": {"url": "https://api.fetchium.com/mcp"}}}