Repository navigation
Capture: smart existing entry selection for LLM prompt #89
Description
Summary
extract_knowledge sends existing entries to the LLM so it can avoid duplicates. Currently sends all keys (truncated to last 50). For large knowledge bases, this wastes prompt tokens on irrelevant entries.
Current behavior
existing_text = "\n".join(f"- {e.key}" for e in existing[-50:])Proposed improvement
Use vector search to find existing entries most similar to the transcript, then only include those in the prompt:
- Embed the transcript (or a summary of it)
query_vector(embedding, limit=20)to get closest existing entries- Send only those 20 entries (keys + short value snippets) to the LLM
This gives the LLM better context for dedup while keeping the prompt small.
Per-model token budgets (future)
Different models have different context windows. Could add per-model config:
phi4-mini: ~4K usable for existing entriesllama3: ~8Kclaude: ~100K
The capture prompt would auto-adjust the number of existing entries based on model capacity.
Acceptance Criteria
- Existing entries sent to capture LLM are filtered by relevance to transcript
- Prompt stays within model capacity
- Capture still detects duplicates effectively
- Falls back to keys-only if embedding unavailable
Discovered during #37 demo testing.
Activity
itdove commentedon Sep 8, 2026
Implemented on branch
89:Verification: focused capture/LLM tests pass (34); full suite has 652 passing tests and 3 unrelated failures caused by the sandbox read-only
/home/itdove/.cache/lorePR-workdir path.itdove commentedon Sep 8, 2026
✅ Session Complete
Project: lore
Goal: #89: Capture: smart existing entry selection for LLM prompt
Time Spent:
0h 14m total
Branch: 89
No significant activity detected in this session.
Session Notes:
Session Notes: itdove-lore-89
JIRA: #89
2026-09-08 11:50
Generated by DevAIFlow
itdove commentedon Sep 8, 2026
Pull request created: #96
itdove commentedon Sep 8, 2026
✅ Session Complete
Project: lore
Goal: #89: Capture: smart existing entry selection for LLM prompt
Time Spent:
0h 15m total
Branch: 89
No significant activity detected in this session.
Session Notes:
Session Notes: itdove-lore-89
JIRA: #89
2026-09-08 11:50
Generated by DevAIFlow
Merge pull request #96 from itdove/89