Repolode
Submit
TencentCloud

TencentCloud/TencentDB-Agent-Memory

A team-level memory hub for AI Agents that turns chats, docs, and code into shared, governed memory assets usable across agent frameworks.

View on GitHubWebsite
Stars
29,631
Forks
2,978
Open issues
772
Last commit
6 hours ago

What is it?

TencentDB Agent Memory addresses the repeated work of re-explaining projects, re-reading documents, and rediscovering workflows each time an AI agent starts a new session. It organizes four kinds of memory assets: Chat Memory for user preferences, facts and decisions; Skills distilled from completed tasks; a Wiki built from product docs and specs; and a CodeGraph indexing code symbols, files, call relationships and impact paths.

Assets are managed in a Memory Hub panel where teams, agents and ownership are defined. Visibility can be private, team, restricted by user, role or agent ACL, or agent-specific, and each agent is equipped with the assets it needs. A Proxy lets existing clients such as Claude Code, Codex, CodeBuddy, Hermes and OpenClaw connect by changing their base URL, without plugins or MCP servers.

The project is deployed with Docker-style startup scripts, and Node.js 22.16 or later is required. Wiki and CodeGraph are processed asynchronously, and CodeGraph currently supports public HTTPS repositories.

Key features

  • 01Chat Memory stores preferences, facts, decisions and interaction history in layers from raw conversation to persona.
  • 02Skills are extracted from conversations and tool calls, with versions, resource files and validation rules.
  • 03Wiki turns docs into structured pages with a link graph; CodeGraph indexes symbols, calls and impact paths.
  • 04Visibility controls range from private to team to ACL-based restricted access, with per-agent asset binding.
  • 05A Proxy connects existing agent clients by changing the base URL, with no plugin or MCP server required.
  • 06Layered retrieval combines BM25, vector search and RRF, with limits on item count and character budget.

Use cases

  • Sharing project architecture docs and code impact analysis across a team of coding agents.
  • Carrying user preferences and past decisions into new sessions of Claude Code or Codex.
  • Turning repeatedly used troubleshooting or release procedures into reusable Skills for specific agents.
  • Importing existing repositories, documents and chat history to bootstrap a new agent team.

Quick start

  1. 1

    Clone the repository

    git clone https://github.com/TencentCloud/TencentDB-Agent-Memory.git
  2. 2

    Enter the deploy directory

    cd TencentDB-Agent-Memory/deploy/global-images
  3. 3

    Copy the env template and fill in LLM parameters

    cp .env.example .env
  4. 4

    Start all services with one command, then open http://localhost:8125

    ./start-all.sh

Things to know

  • Requires two sets of LLM parameters (memory group and proxy group) to be filled in before it runs.
  • The MongoDB storage backend is experimental and off by default.
  • The license is listed as NOASSERTION in metadata, so verify the license terms yourself despite the README badge showing MIT.

Topics

#agent#ai-agent#long-term-memory#llm#vector-search#embedding#local-first#code-graph#memory#openclaw-plugin

FAQ

What is TencentDB-Agent-Memory?
A team-level memory hub for AI Agents that turns chats, docs, and code into shared, governed memory assets usable across agent frameworks. TencentDB Agent Memory addresses the repeated work of re-explaining projects, re-reading documents, and rediscovering workflows each time an AI agent starts a new session. It organizes four kinds of memory assets: Chat Memory for user preferences, facts and decisions; Skills distilled from completed tasks; a Wiki built from product docs and specs; and a CodeGraph indexing code symbols, files, call relationships and impact paths.
Is TencentDB-Agent-Memory free to use? What license does it have?
TencentDB-Agent-Memory is open source under the Other license. Check the license terms before using it in commercial products.
Is TencentDB-Agent-Memory still maintained?
Actively maintained. The last commit to TencentDB-Agent-Memory was on Oct 19, 2026.
What can I use TencentDB-Agent-Memory for?
Sharing project architecture docs and code impact analysis across a team of coding agents. Carrying user preferences and past decisions into new sessions of Claude Code or Codex. Turning repeatedly used troubleshooting or release procedures into reusable Skills for specific agents. Importing existing repositories, documents and chat history to bootstrap a new agent team.
How do I install TencentDB-Agent-Memory?
Clone the repository: git clone https://github.com/TencentCloud/TencentDB-Agent-Memory.git Enter the deploy directory: cd TencentDB-Agent-Memory/deploy/global-images Copy the env template and fill in LLM parameters: cp .env.example .env Start all services with one command, then open http://localhost:8125: ./start-all.sh
Where can I get TencentDB-Agent-Memory?
The source code and installation instructions are on GitHub: https://github.com/TencentCloud/TencentDB-Agent-Memory

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects