Family documents
StaleStrong freshnessForwarded school, camp, and sports documents that ground Family Manager proposals.
- Documents
- 86
- Chunks
- 642
- Tokens
- 412,300
- Storage
- 17.5 MB
- Embedded
- 97%
- Updated
- 4d ago
strong
Retrieval precision
Canned queries return on-topic chunks.
3
Stale documents
Three docs past their freshness window.
1
Unembedded docs
One document is awaiting embedding.
0
Orphan chunks
No chunks without a parent document.
healthy
Source coverage
All forwarded source types represented.
Inventory
Documents and their chunk metadata. Open any document for source type, embedding status, token counts, and chunk ids, never raw content or vectors.
Retrieval
Canned query examples show which chunks were returned, the coarse similarity score, and why each chunk matched.
“When does summer camp start and end?”
Run by FM Doc Ingest
- 0.91chk_camp_1 · doc_camp_flyer
Mentions session start and end dates.
- 0.78chk_camp_3 · doc_camp_flyer
Lists drop-off and pick-up windows.
“What are the school term break dates?”
Run by Family Manager
- 0.84chk_cal_1 · doc_school_calendar
Contains term boundary dates.
These are curated, canned queries. Arbitrary querying over private documents is intentionally disabled.
Learning
Constrained autolearn: candidates from feedback pass eval and reviewer gates before promotion, with demotion and rollback when they regress.
Autolearn is constrained: feedback creates candidates and promotion gates, it does not silently mutate production. Reviewer decisions and rollbacks are shown below.
- cand_rsvp_patternFM Doc IngestPromotedEval: StrongReviewer: approvedSignals:parent 11reviewer 2decided 4d ago
- cand_carpool_hintFM Doc IngestIn reviewEval: ModerateReviewer: pendingSignals:parent 4
- cand_bad_splitFM Doc IngestRolled backEval: WeakReviewer: rejectedSignals:parent 3reviewer 1decided 4d ago
Freshness timeline
Recent writes, re-embeds, retrieval regressions, and candidate promotions and rollbacks.
KB write×1
4d ago
Candidate promoted×1
4d ago
Re-embed×7
4d ago
Candidate rolled back×1
4d ago
Retrieval regression×2
8d ago