Daedra — Self-contained Web Search MCP Server

Multi-Backend Fallback

14 unkeyed backends run each query. Results merge by relevance. Mwmbl → Brave → Marginalia → Bing RSS → Bing → Google News → HN → Google → Wikipedia → StackOverflow → GitHub → Wiby → DDG Instant → DuckDuckGo HTML.

Search Reliability

Per-backend circuit breakers and governor keyed rate limits. Transient errors get classified exponential-backoff retry; CAPTCHA and rate limits fail fast.

Content Extraction

visit_page: HTML via dom_smoothie Readability, PDF via infer + pdf-inspector, Office documents via anydoc. Responses classified as Html, Pdf, Document, Text, or Binary before Markdown conversion.

Works From Any IP

Mwmbl, Marginalia, Wikipedia, StackExchange, GitHub, Wiby, and DDG Instant always work — even from datacenter/VPS IPs where Google, Bing, and DDG HTML are CAPTCHA-blocked.

MCP Server

Drop-in MCP for Claude, Cursor, pawan, or any MCP client. STDIO and SSE transports. Tools: web_search, visit_page, crawl_site (search_duckduckgo alias).

Pure Rust

No Docker. No Python. Single binary. Async I/O, connection pooling, moka response caching. Daedra IS the search infrastructure.

Search Backends

Backend	Type	API Key	Works from VPS?
Mwmbl	Public JSON API	None	Always
Brave	HTML scraping	None	Sometimes
Marginalia	Public JSON API	None / `MARGINALIA_API_KEY`	Always
Bing RSS	RSS	None	Always
Bing	HTML scraping	None	Sometimes
Google News	News RSS	None	Always
Hacker News	Algolia JSON	None	Always
Google	HTML scraping	None	Rarely
Wikipedia	OpenSearch API	None	Always
StackExchange	Public API	None	Always
GitHub	Public API	None / `GITHUB_TOKEN`	Always
Wiby	Indie web search	None	Always
DDG Instant	Knowledge graph API	None	Always
DuckDuckGo	HTML scraping	None	Rarely

Quick Start

MCP Server

{
  "mcpServers": {
    "daedra": {
      "command": "daedra",
      "args": ["serve", "--transport", "stdio", "--quiet"]
    }
  }
}

CLI

daedra search "rust async runtime" --num-results 5
daedra fetch https://rust-lang.org
daedra check