Hippocampus
Hippocampus solves the most annoying thing about daily AI agent use: starting every session from scratch.
Added
1 month ago
About
Hippocampus solves the most annoying thing about daily AI agent use: starting every session from scratch. It reads your agent's existing daily notes, synthesises a rolling 14-day briefing, and injects that context automatically before each new session — no manual recap required. Persistent patterns get promoted to a permanent memory layer. Free, open source, and installs with one command.
person_check
Best For
OpenClaw users running ongoing projects where continuity matters: long-form research, multi-week coding projects, or any workflow where re-explaining context at the start of each session costs real time.
Pros & Cons
check_circle
Pros
- check Zero friction setup: reads daily notes your agent already writes, no new data structures to maintain
- check Rolling briefing keeps context window usage lean (3,000–5,000 characters) while covering two weeks of history
- check Two-tier memory (rolling briefing + permanent layer) balances recency with long-term retention
- check MIT licensed with no API key or account required — genuinely free to use
- check Works with agent fleets, not just solo setups
cancel
Cons
- close 14-day rolling window means context older than two weeks disappears unless explicitly promoted
- close Synthesis is LLM-based, so output quality varies and isn't deterministic
- close Requires manual cron job setup for automated scheduling — no built-in scheduler
- close Depends on agents consistently writing daily notes; sporadic note-takers get less benefit
More Memory
Other tools in the same category.
