PROJECT: memory-compactor

Aim

Compact noisy conversational memory stretches (especially debug sessions) into durable resolved summaries while preserving recall quality.

Current implementation

HTTP client for custom mem0 server endpoints:
- POST /memories/all
- POST /memories
- DELETE /memory/{id}
temporal clustering (gap-minutes)
debug-arc heuristic and compact summary writer
dry-run reporting

Operating mode

default dry-run
--apply to mutate memory state
bounded deletions with --max-deletes

Design notes

keep newest item in each compacted cluster as guardrail
write compacted summary before deleting old entries
annotate summaries with provenance metadata (compacted_from_ids)

Next steps

Add rollback snapshot file before apply
Add explicit per-user allowlist
Add cron integration helper (OpenClaw cron payload template)
Add quality checks pre/post compaction