TencentDB Agent Memory
About
A team-level memory hub for AI Agents — turning conversations, docs, and code into four reusable memory assets (Chat Memory, Skill, LLM-Wiki, Code-Graph) that are governed, shared, and equipped across agents and frameworks. Supports OpenClaw, Hermes, Claude Code, CodeBuddy and more.
Our Verdict
TencentDB Agent Memory is a serious contender in the young agent-memory space: it turns the "memory" idea from a nice-to-have into a governed, shareable team asset with layered distillation (L0-L3), permission-aware sharing, and framework-agnostic design. The one-command deployment and local panel lower the barrier considerably. Still, it is a Beta with a notable open-issue count, asynchronous indexing, and a CodeGraph that favors public repos — early adopters should budget time for setup and iteration.
Pricing
Use Cases
- Giving a team of AI agents shared long-term memory across sessions and frameworks
- Turning an existing codebase and docs into searchable Wiki and CodeGraph assets
- Letting new agents cold-start from accumulated team experience instead of learning from scratch
Pros
- Converts conversations, docs, and code into four reusable memory assets automatically
- L0-L3 layered memory distillation with BM25 + vector retrieval + RRF fallback
- Privacy by design: private-by-default sharing with four visibility levels
- Framework-agnostic — supports OpenClaw, Hermes, Claude Code, CodeBuddy and SDK
- One-command deployment with a local management panel
Cons
- Project is in early stages (Team Memory Beta) with 553 open issues
- Requires configuring two sets of LLM parameters and running local services
- Wiki and CodeGraph are built asynchronously — need processing time to become ready
- CodeGraph currently prioritizes public HTTPS repos; private repo/SSH support still maturing
- Fully automated memory routing is still under iteration
Installation
git clone https://github.com/TencentCloud/TencentDB-Agent-Memory.git && cd TencentDB-Agent-Memory/deploy/global-images && cp .env.example .env && ./start-all.shFrequently Asked Questions
What is TencentDB Agent Memory?
It is Tencent Cloud's open-source team-level memory hub for AI Agents. The core idea: project context already explained, documents already read, and workflows already proven should not be re-discovered from scratch by every agent. It distills conversations, docs, and code into four memory assets — Chat Memory (L0-L3 layered distillation), a Skill library, LLM Wiki (document knowledge graph), and CodeGraph — managed by a Memory Hub and shared with team agents under permission control.
How do I install and deploy it?
One command brings up the full trio (memory-core + memory-hub + proxy): clone the repo, enter deploy/global-images, copy .env.example to .env and fill in two sets of LLM parameters, then run ./start-all.sh. When finished it prints a one-liner you can paste directly into Claude Code; the management panel is at http://localhost:8125. Full docs in INSTALL.md.
Is it free? Which agent frameworks are supported?
Completely free and open source (MIT). Officially supports OpenClaw, Hermes, Claude Code, CodeBuddy and SDK integration; broader cross-framework migration is on the roadmap. Privacy by design: new Chat Memory and Skills are private by default, sharing is an explicit action — visibility levels span private / team / restricted / agent.