What it does
Vestige is a local-first memory layer for AI agents. It uses FSRS-6 scheduling and SQLite storage, with smart ingest, portable sync, an embedded dashboard and optional hooks. The project also describes a deterministic gate that checks agent actions and keeps a receipt of each one.
Use cases
- 01Keep project knowledge across agent sessions
- 02Review a record of actions an agent took
- 03Sync agent memory between machines