Search Fragments
Server Details
Search Fragments is the gate before your agent answers. When a user half-remembers something your agent can't confidently place, one call returns a verdict: ORIGIN_OF_TRUTH (found and evidence-checked, with sources), PARTIAL_ORIGIN (leads worth following, not an answer) or NOT_REACHABLE (our search reached nothing; not a claim it can't be found). Your agent answers and cites, keeps looking, or says it can't place it, instead of inventing. Free, no signup. Also: verify_claim (check a specific factual claim against current sources) and submit_resolution_feedback.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- Repository
- Searchfragments/search-fragments-mcp
- GitHub Stars
- 0
- ChatGPT
- –
- Claude
- Search Fragments
TDQS
Scored across 4 tools
The four tools split cleanly along a pipeline: cheap routing (gate_probe), full identification (resolve_fragment), feedback (submit_resolution_feedback), and fact-checking (verify_claim). The only friction is that gate_probe can itself return an answer via QUICK_LOOKUP, so its boundary with resolve_fragment depends on the agent respecting the stated call-order rather than on an inherent resource difference.
All names are snake_case verb_noun (gate_probe, resolve_fragment, submit_resolution_feedback, verify_claim) with a clear action-first convention. The routing verdict constants are uppercase and live in descriptions rather than names, so they don't pollute the naming scheme.
Four tools is lean but defensible for a narrow fragment-identification service: a probe, a resolver, a feedback sink, and an adjacent fact-checker. Each has a distinct role, though verify_claim is arguably a separate capability grafted onto the core loop.
The probe → resolve → feedback lifecycle is closed, and PARTIAL_ORIGIN leads can be pursued by re-calling resolve_fragment, so there are no hard dead ends. Minor gaps remain: no way to expand a shortlist item into full source detail, and no read-back of submitted feedback.
Available Tools
4 toolsgate_probeGate ProbeAInspect
The gate before your agent answers. Call this FIRST for any question, before spending effort: one cheap look returns a routing verdict in a few seconds.
QUICK_LOOKUP: one page names the answer, with a quote copied from that page. Check the quote, then answer.
NEEDS_FULL_JOB: worth a full search. Call resolve_fragment with the user's words.
NOT_WORTH_IT: SF's one look found no page naming an answer, so a full search is unlikely to pay off (about 1 in 10 of these is still findable). Tell the user you cannot place it rather than guessing.
PROBE_UNAVAILABLE: the look failed. Treat as NEEDS_FULL_JOB.
Every outcome is a routing estimate about SF's own look, never a claim that the thing cannot be found.
Inputs: fragment (required).
resolve_fragmentResolve FragmentAInspect
Returns a verdict first, so your agent knows whether to answer, keep looking or say it can't place it: ORIGIN_OF_TRUTH (found, evidence checked, with sources), PARTIAL_ORIGIN (leads to follow, not an answer) or NOT_REACHABLE (SF's search reached nothing; not a claim the thing can't be found). Pieces together the name of a film, TV or radio show, song, book, game, quote, person, place, event or research study from the fragments a user remembers: a scene, a line, a face, roughly when. Pass every fragment in the user's own words, even uncertain ones. It searches many sources and fits them together. A shortlist or an honest "can't place it" is still useful. Use it when your own search couldn't place it. Not for known titles, trivia, recommendations or checking claims (use verify_claim). When the user says whether this helped, report it with submit_resolution_feedback.
Inputs: fragment (required), skip_gate. Returns: pool, leads, outcome, summary, +10 more
submit_resolution_feedbackSubmit Resolution FeedbackAInspect
After the user reacts to a resolve_fragment result, call this with its resolution_id. accepted: it is what they meant. partial: it helped but is not quite it. rejected: it is not what they meant. If they recognised an item in the shortlist or leads, send picked_rank (1 = first item) or picked_url; if nothing fitted, send none_fitted: true. corrected_target and new_clue are optional and are stored as feedback only; they never change graph facts.
Inputs: status (required), resolution_id (required), note, new_clue, +4 more. Returns: status, recorded, resolution_id
verify_claimVerify ClaimAIdempotentInspect
Checks whether current web sources support a specific factual assertion. Takes a claim as a natural-language statement; returns a verdict with cited evidence and stated limits.
Verdict set:
"supported" — multiple independent sources directly confirm the specific claimed detail
"partially_supported" — sources confirm the entity and domain; specific detail is pointed to but not directly stated
"insufficient_evidence" — sources don't address the specific claim; fires freely, including when the entity is well-known but the specific detail is undocumented
"unsupported" — a credible source explicitly contradicts the claim
All verdicts are DECIDE-BY-EYE. A "supported" verdict means current web sources confirm it — not that it is true. Evidence may be incomplete, biased, or outdated. The stated_limits field is always present and identifies what the evidence cannot confirm.
Good input shape (specific, checkable assertions):
"Werner Herzog dragged a full-size steamship over a hill during the filming of Fitzcarraldo"
"Glenn Gould stopped giving live concerts in 1964"
"The Backrooms photograph originated on 4chan"
Checks one specific factual statement. Not for identifying half-remembered works.
Not this shape:
"there's a documentary about a filmmaker dragging a boat over a mountain"
"a musician famous for stopping performing"
Inputs: claim (required). Returns: verdict, evidence, stated_limits
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Cited web evidence and fact checking for AI agents. $0.01 per call, no API key.
Check whether a real-world fact can be verified before an agent acts on it. Free, no auth.
Real-time fact-check, citation verification, and source-freshness for AI agents.
Free Perplexity citation check for your domain and prompt. 5 calls per caller per day.
Related MCP Servers
- AlicenseNot gradedqualityDmaintenanceWeb search, clean page reading & one-call research dossiers for AI agents. No API key — your agent does the synthesis.81 npmMIT

jevbook MCPofficial
AlicenseNot gradedqualityCmaintenanceEnables agents to fact-check claims, conduct verified research, make typed decisions with calibrated probabilities, and scan token risks through a hosted server with no API key required.MIT- AlicenseAqualityBmaintenanceEvidence-backed web research for AI agents. Real-time search with cited claims, confidence scores, and compare mode showing raw LLM hallucination vs evidence-backed answers.520Apache 2.0
- AlicenseNot gradedqualityBmaintenanceEnables AI agents to search the live web, fetch clean page content, extract schema-driven structured JSON, check link liveness, detect page changes, run recurring monitors, and verify claims against sources. It offers a free per-IP tier with no API key, plus pay-per-call or prepaid access for higher volume and LLM-backed tools.MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.