Pin the source
The document and revision determine what may be cited. Re-ingestion is not proof the upstream source stayed fresh.
Source code ↗A source-to-proof system
One pinned document. One technical question. The answer is only as useful as the evidence the system can recover and let you inspect.
How should running work finish while jobs are queued?
From source to answer
Each layer has a reason to exist, a cost worth measuring and a failure condition that should be visible.
The document and revision determine what may be cited. Re-ingestion is not proof the upstream source stayed fresh.
Source code ↗Chunk, rank and pack passages within a limited context budget. Compare against BM25 before trusting sophistication.
Source code ↗Missing supporting context should produce an explicit limitation. A citation link alone does not prove entailment.
Source code ↗Extractive or generated output is labeled honestly, with citations and diagnostic traces available for review.
Source code ↗What the experiment actually shows
Frozen 15-question, single-README evidence-recall diagnostic; approximate 1,400-token context budget. This does not measure generated-answer accuracy.
Required-evidence recall after budget-aware passage trimming.
Same recall as plain BM25 on this testUnsupported probes refused by the evidence gate in this narrow diagnostic.
Not a general hallucination or safety scoreThe experiment started at 36.7% recall with incorrect document-level deduplication. The fix recovered more required passages; the next step is independent corpora and human-judged answers. Read the evaluation and raw runs.
Trust boundaries
Only when the relevant question, corpus, source revisions, settings, policy and model path still match. A cached answer can be fast but stale; quality tests deliberately bypass the cache. Source identity and model context are different types of state.
Not established. The current visible graph links documents, headings, shared terms and feedback. It is a diagnostic, not proof of semantic relationships or better retrieval. An independent benchmark must justify graph-based execution.
No. The host is explicitly in an ephemeral, read-only demo posture. Custom uploads, visitor model keys and destructive mutation paths are refused server-side. Private workspaces and durable storage require separate authenticated infrastructure.
The public host has no configured server generation key. Its extractive response quotes source material; it does not pretend a model synthesized the conclusion. This makes the lack of model availability visible rather than hiding it behind sample prose.