A free, unlimited web search MCP server for LLM agents using DuckDuckGo
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"ddg-search-mcp","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# DuckDuckGo Search MCP Server A free, unlimited web search MCP server for LLM agents. Uses DuckDuckGo to search the web and returns full extracted page content. ## How it works 1. Your LLM agent calls the `web_search` tool with a query 2. The server searches DuckDuckGo and gets the top results 3. It concurrently fetches each result page and extracts the main content, stripping ads, navigation, and boilerplate 4. Clean, structured Markdown is returned — ready for the LLM to reason over and cite Results are sorted with full page extracts first, falling back to snippets for pages that can't be fetched (paywalls, bot blocks, etc.). A summary header tells the agent how many full extracts vs snippets were returned. --- ## Requirements - [Claude desktop app](https://claude.ai/download) - [uv](https://docs.astral.sh/uv/getting-started/installation/) installed --- ## Installation ```bash claude mcp add web-search uvx ddg-search-mcp ``` That's it. No virtual environments, no pip installs, no API keys. `uvx` handles everything automatically in an isolated environment. Verify it's connected: ```bash claude mcp list ``` You should see `web-search` with a ✓ connected status. --- ## Usage Once installed, just talk to Claude naturally: > *"Search for the latest news on AI regulation"* > *"Find recent benchmarks comparing Claude and GPT-4"* > *"Look up how to configure a Next.js app router"* Claude will automatically call the `web_search` tool when it needs current information from the web. You can also explicitly ask: > *"Use web search to find..."* ### Tool parameters | Parameter | Type | Default | Description | |-----------|------|---------|-------------| | `query` | string | required | Natural language search query | | `num_results` | int | 5 | Number of results to fetch (1–10) | --- ## Performance - Pages are fetched **concurrently** — 5 results take roughly the same time as 1 - Pages that take longer than **8 seconds** are dropped automatically - …
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.