Fully-local long-term memory for AI agents built on two pillars: layered long-term memory (a semantic pyramid L0 Conversation -> L1 Atom -> L2 Scenario -> L3 Persona) and symbolic short-term memory that offloads verbose tool logs to files while keeping a compact Mermaid 'canvas' in context. Distributed as a TypeScript/npm package (@tencentdb-agent-memory/memory-tencentdb); integrates with OpenClaw and Hermes via plugins and exposes a REST Gateway API at port 8420 for use from any HTTP-capable agent framework.
- Storage
- Dual-layer heterogeneous storage: bottom layers (atomic facts, logs, raw tool outputs as refs/*.md, step summaries as jsonl) persist in SQLite + sqlite-vec for full-text and vector retrieval; top layers (personas, scenes, Mermaid canvases) are stored as human-readable Markdown. A deterministic node_id drill-down path links each abstraction back to ground-truth evidence. Optional Tencent Cloud Vector Database (TCVDB) backend available via migration script.
- Retrieval
- Hybrid retrieval: BM25 + vector + RRF (ranked reciprocal fusion) over the stored layers, supporting both keyword and semantic recall. Hierarchical recall: the agent attends to the top-layer Persona/canvas in context and drills down to mid-layer indexes (Scenario/jsonl) and bottom-layer raw text (L0/refs) via node_id when detail is needed.
- Self-host
- Self-host: moderate
- License
- MIT
- Pricing
- Open-source (MIT), free to self-host; fully local with zero external API dependencies. · Free / OSS
- GitHub stars
- 19,196
- Last release
- 2026-08-03
- Last commit
- 2026-08-10
- First catalogued
- 2026-06-28
Strengths
- Layered design (L0->L3 semantic pyramid) avoids flat vector dumps and preserves a traceable drill-down path back to source evidence
- Symbolic short-term memory (Mermaid canvas + history offloading) targets token reduction on long-horizon tasks
- Hybrid retrieval (BM25 + vector + RRF) supports both keyword and semantic recall within the same pipeline
- REST Gateway API at port 8420 makes the memory layer callable from any HTTP-capable agent framework, not just OpenClaw or Hermes
- Fully local with no external API dependency; backed by Tencent
Watch out
- Young project (created 2026-04) — the API and storage surface may still be moving
- License: README badge says MIT and package.json confirms MIT, though GitHub auto-classifies it 'NOASSERTION' (non-standard header in LICENSE file)
- Requires Node.js 22.16 or later — strict minimum version barrier for npm users
- README carries self-reported benchmark numbers (WideSearch, SWE-bench, AA-LCR, PersonaMem) — record those via harvest-benchmarks with selfReported:true, not on this card
Best for
- Long-horizon agent tasks needing token-efficient, fully-local memory with traceable layered recall
- Agents running on OpenClaw or Hermes wanting zero-config long-term memory with no external API
- Teams that want a standalone REST Gateway memory sidecar callable from any agent framework
How it integrates
Benchmark results
No sourced results yet.
Sources
- Monthly card re-verification — GitHub API repo metadata: 10,029 stars, GitHub reports no auto-detectable SPDX license (card license recorded from the repo's own LICENSE file), archived=false, last commit on the default branch 2026-07-29 (third-party)
- TencentDB Agent Memory README (vendor)
- GitHub API repo metadata (stars, release) (third-party)
- LICENSE (MIT, Copyright Tencent) (vendor)
- npm package @tencentdb-agent-memory/memory-tencentdb (v0.3.6) (vendor)
- GitHub release v0.3.6 (2026-05-28) changelog (vendor)
- package.json — integrations, engines, bin targets (vendor)
- TencentDB Agent Memory README + CHANGELOG (monthly re-verify) — architecture, storage, retrieval, OpenClaw/Hermes/Gateway integrations unchanged; npm package now at v1.0.1 with GH release tags v1.0.0/v1.0.1/v2.0.0-beta.1 all marked prerelease (stable 'latest' release remains v0.3.6, matching lastReleaseDate) (vendor)
- GitHub API license endpoint — still returns NOASSERTION (non-standard LICENSE header); LICENSE file text confirms MIT (third-party)
Last verified 2026-08-01 · updated by refresh-framework-cards