This release adds 3 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Topics
+4 more
Summary
AI summaryIntroduced a self‑organizing topic tree with hierarchical clustering, LLM naming, and a new 4‑segment resume format.
Full changelog
What's New
Topiary: Self-Organizing Topic Tree
- Hierarchical memory clustering via k-means with L2-normalized centroids
- Background worker with 5s debounce, triggered by embed flush, consolidation, and memory deletion
- LLM-powered topic naming (batch 30, dirty flag tracking)
- Cached in DB for zero-latency resume reads
4-Segment Resume
- New format: Core → Recent → Topics → Triggers
- Topics section: named topic index with drill-down IDs
- Triggers section: aggregated from all layers
API
POST /topic— retrieve full memories for specific topics by ID- Resume JSON includes
topicsandtriggersfields
Web UI
- Resume page adapted to 4-segment format
- New Topics page with click-to-expand drill-down
- LLM Usage: added naming and embed_queue descriptions
Bug Fixes
- Fix triggers not showing in resume (JSON array parse bug)
- Fix Topics header showing only Working+Buffer counts (now includes Core)
- Trigger topiary rebuild on memory delete (single + batch)
- Remove redundant inline trigger tags from Core section
MCP
[email protected]: addedengram_topictool
Full Changelog: https://github.com/kael-bit/engram-rs/compare/v0.12.0...v0.12.1
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About kael-bit/engram-rs
Hierarchical memory engine for AI agents with automatic decay, promotion, semantic dedup, and self-organizing topic tree. Single Rust binary, zero external dependencies.
Related context
Related tools
Beta — feedback welcome: [email protected]