TencentCloud/TencentDB-Agent-Memory
An open-source memory hub turns agent conversations, workflows, docs, and code into governed assets that can be reused across sessions and roles, reducing repeated project setup.
TencentDB Agent Memory packages prior work into **four memory assets**: Chat Memory, Skills, Wiki, and CodeGraph. The hub adds ownership, versions, agent bindings, and private, team, or ACL-based access.
Use it to give coding and review agents different context loadouts instead of one global prompt. Memory is distilled through **L0–L3**, while detailed recall combines **BM25, vector retrieval, and RRF** under context-size and timeout limits.
TencentDB Agent Memory packages prior work into **four memory assets**: Chat Memory, Skills, Wiki, and CodeGraph. The hub adds ownership, versions, agent bindings, and private, team, or ACL-based access. Use it to give coding and review agents different context loadouts instead of one global prompt. Memory is distilled through **L0–L3**, while detailed recall combines **BM25, vector retrieval, and RRF** under context-size and timeout limits. The Team Memory release is still a beta. Wiki and CodeGraph indexing is asynchronous, private-repository support and automatic routing remain incomplete, and current compatibility is limited to **OpenClaw, Hermes, and SDK integrations**.
This makes the proposed shared context layer more concrete by packaging governed, versioned assets with per-agent bindings, tiered distillation, and hybrid retrieval. It also exposes the gap between architecture and production readiness: routing, repository coverage, indexing latency, compatibility, and recipient-level access still constrain a portable team-memory layer.