Trust
AI Memory Provenance: Why Source IDs Matter
Understand why AI memory needs source trails, supersession, and review before old context steers new work.
Article packet
Concepts
Users who need memory they can trust, inspect, and correct
5 min read
01
Provenance answers where a memory came from.
02
Supersession answers what changed.
03
Review answers whether a record should guide future context.
01
Quick answer
AI memory provenance is the trail that connects a remembered fact to source memory IDs, page source lists, source-agent metadata, and supersession or revision state. Session logs are useful context, but they are not per-fact provenance.
Origin keeps source_memory_ids with page records, exposes source memories through page-source APIs, tracks revisions, and writes local git history for readable artifacts so memory work stays inspectable.
02
When this problem appears
Without provenance, memory becomes a black box. An assistant may retrieve a stale claim, but the user cannot tell when it was created, what supported it, or whether it was superseded.
03
Prefer traceable memory
Make memory useful without making it mysterious.
- Capture why a fact matters.
- Use corrections when facts change.
- Use review for low-confidence or conflicting records.
- Use source-backed pages for synthesis.
- Use source IDs, page sources, revisions, review, and forget flows for atomic memory inspection.
- Use local git history for readable artifact changes.
04
What to check next
Provenance does not make every memory true. It makes claims inspectable so stale or wrong records can be corrected.
Try the local memory loop
Install Origin, connect your AI client, and verify that capture, recall, and handoff work on your machine.
FAQ