Two tools discovered
The server exposes search(query, max_results, region) and fetch_content(url, ...), a compact surface for web retrieval.
evidence/trust5/2026-07-03-mcp-pilot.json
MCP serverPublished
DuckDuckGo Search MCP Server
The DuckDuckGo Search MCP server passed a real smoke test: a live search returned relevant ranked results and page fetch returned readable content with character accounting. The score is moderate because the server scrapes DuckDuckGo without an official API, so reliability and continued operation depend on a third party that has not agreed to serve it.
Independent trust badge
The visible trust mark for this verdict.
Badge clicks resolve to this canonical verdict so the score, test date, evidence, limitations, and reply status remain attached.
Embed
Show this badge on your site
[](https://silentcritique.com/verdicts/duckduckgo-mcp)Markdown works in GitHub READMEs. The badge always links back to this verdict.
Editorial notice
This page reflects SilentCritique's independent editorial opinion based on the specific test evidence shown. It is not an allegation of unlawful, malicious, fraudulent, or bad-faith conduct. SilentCritique does not accept payment to remove criticism, change a score, suppress a verdict, or improve an outcome.
Claim tested
Can the public DuckDuckGo MCP server run a live web search and fetch page content over MCP without any API key?
Evaluator panel
Protocol harnessReliability reviewerOperator skeptic
Evidence reviewed
Live search returned relevant results
The query returned 3 results led by modelcontextprotocol.io, each with URL and summary.
evidence/trust5/2026-07-03-mcp-pilot.json
Page fetch returned readable content
fetch_content on example.com returned extracted text with an explicit character-range note (139 of 139 characters shown).
evidence/trust5/2026-07-03-mcp-pilot.json
Test setup
- Started duckduckgo-mcp-server==0.5.0 over MCP stdio via uvx.
- Used the official MCP client SDK to list tools, search for "model context protocol" with max_results 3, and fetch https://example.com/.
- Stored the full tool-call evidence in evidence/trust5/2026-07-03-mcp-pilot.json.
Strengths
- Worked with zero configuration and no API key in the tested run.
- Result formatting is structured enough for an agent to cite URLs.
- Content fetch reports how much of the page was returned, which limits silent truncation.
Failure modes
- Scraping-based search can be rate limited or blocked upstream without notice, and results may differ from an official API.
- Fetched page content enters the agent context unfiltered, which is a prompt-injection path the host must handle.
- One search and one fetch were probed; pagination, regions, and error paths were not.
What would improve the score
- Document rate-limit behavior and failure modes when DuckDuckGo blocks requests.
- Offer an option to sanitize or annotate fetched content before it reaches the model context.
Limitations
- This was an unsolicited smoke test of the public package, not a source audit.
- Live results depend on DuckDuckGo availability at test time.
Visible dissent
- The reliability reviewer withheld points because the upstream is unofficial and can break silently.
- The harness reviewer noted the tested happy path worked exactly as documented.
Right of reply
No vendor reply has been requested or published as of 2026-07-03. SilentCritique will publish factual corrections or a right of reply through the corrections process.
Do you build DuckDuckGo Search MCP Server?
Claim this verdict to publish a reply, correct factual errors, or request a re-test after you ship fixes. Replies are published verbatim next to the score.
Claim this verdictMethodology matters
Scores are only meaningful when the rubric, date, evidence, and dissent are visible.
Read methodology