TencentDB Agent Memory is 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.
A brain that remembers people and context. Retains preferences, facts, decisions, and interaction history. Each Agent automatically gets its own memory when created — no need to re-introduce yourself next time.
A skill library that accumulates expertise. Extract and manage reusable Skills from conversations and tool calls. A Skill isn't just a prompt snippet; it has versions, resource files, trigger boundaries, and validation rules.
A knowledge map that reads both docs and code. Turns product docs, design specs, and ops runbooks into structured pages with a link graph. Agents can search, read, and inspect relationships.
Indexes code symbols, files, call relationships, and impact paths. Agents can search, inspect callers/callees, and perform impact analysis before modifying code. Tells them "changing this might affect those."
# Clone and start all three services in one go
git clone https://github.com/Tencent/TencentDB-Agent-Memory.git
cd TencentDB-Agent-Memory/deploy/global-images
cp .env.example .env
$EDITOR .env # Fill in two sets of LLM parameters
./start-all.sh # Launch everything with one command
Open the panel: http://localhost:8125
Let experience accumulate, flow, and pass on to the next Agent
Import existing documents, codebases, and conversation sessions. Start from existing experience
New Chat Memory and Skills are private by default. Sharing is an explicit action
Bind different memory assets to different Agents; adjust priority and usage mode
Build Wiki and CodeGraph; monitor processing status and asset metadata
Switch between private, team, and ACL-based access; revoke sharing when needed
| Benchmark | Without TencentDB Agent Memory | With it enabled | Relative improvement |
|---|---|---|---|
| PersonaMem | 48% | 76% | +59% |
PersonaMem tests whether an Agent can correctly understand and apply user information after extended interactions.
We welcome contributions of all kinds — bug reports, feature suggestions, documentation fixes, benchmark reproductions, ecosystem integrations, or pull requests. Agent memory is far from settled, and we hope to build it together with the community.