engrim vs MemMolt
Local SQLite episodic memory shared across Antigravity, Claude Code, Cursor, Windsurf and Codex: hooks/MCP capture decisions, FTS5+model2vec hybrid recall reloads ~1k tokens per session. — versus — Structured long-term memory over MCP: an enforced bucket-thread-memo hierarchy in one SQLite file, hybrid FTS5 + vector search fused with RRF, local embeddings.
Same single-SQLite-file, FTS5-hybrid, MCP-served memory design; memmolt enforces a bucket-thread-memo hierarchy, engrim focuses on multi-harness hooks and a bounded memory pack.
| engrim | MemMolt | |
|---|---|---|
| Stars | 222 | 4 |
| Forks | 12 | 0 |
| Language | Python | JavaScript |
| License | MIT | MIT |
| Last activity | yesterday | 5 months ago |
| Topics | memory, coding | memory |
| Curated connections | 5 | 6 |
engrim — the curator's take
Use it when you actually switch harnesses mid-project — Gemini in Antigravity in the morning, Claude Code in the afternoon — and want one memory both write to and read from, with `origin_agent` provenance on every entry. `engrim setup` auto-wires hooks, MCP and status lines for whatever it finds under ~/. The retrieval is cheap and offline (FTS5 + static model2vec embeddings, no API), and the deliberate 4k-char 'memory pack' cap keeps it from becoming a second context window. Don't pick it for cross-project or team memory (it is per-project, per-machine), or if you want raw transcript capture rather than curated decisions/facts/feedback. Small project (~200 stars, one maintainer) — expect rough edges.
MemMolt — the curator's take
The anti-sprawl memory play: a forced 3-level hierarchy the agent can't turn into a jungle, with ~10ms hybrid search from a single SQLite file and zero cloud calls. Young and tiny (4 stars) — the schema idea is worth studying even if you don't adopt it. Skip if you want auto-capture; this is deliberate, curated memory.