RepoRadar
Leave a Tip
Item detail

TencentCloud/TencentDB-Agent-Memory

TencentDB Agent Memory is a MIT-licensed memory plugin for AI agents that delivers symbolic short-term memory (compact Mermaid symbols instead of bloated tool logs) and layered long-term memory (structured personas + scenes instead of flat vector piles). It ships as a Node package (`@tencentdb-agent-memory/memory-tencentdb`), plugs into OpenClaw and Hermes Agent, and ships benchmark numbers: 61.38

Open source →Back to radar
Score7.9
Popularity80.0
Risknone
TierSilver
Score breakdown
Usefulness8.0
Novelty9.0
Momentum8.0
Maturity7.4
Open-source/build8.4
Evidence7.2
Workflow potential9.4
Setup ease8.8

Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.

Why it matters

Useful for agent teams burning tokens on long-horizon tasks and tired of flat-vector memory that loses structure. npm install the plugin, point OpenClaw or Hermes at it, and watch the symbolic + layered pipeline cut tool-log token usage while raising pass rate on long-horizon benchmarks like WideSearch and PersonaMem.

Who should use it

agent teams burning tokens on long-horizon tasks who need to cut tool-log overhead without losing recallteams running OpenClaw or Hermes Agent who want a drop-in memory plugin instead of writing their own Mem0/Zep layerdevelopers building personal-assistant agents who need per-user personas and scenes, not a flat vector pileresearchers who want to benchmark symbolic + layered memory against their own flat-vector baselineTencent Cloud users who want a Tencent-blessed memory layer that plugs into the rest of the OpenClaw / Hermes stack

Who should skip it

Skip if the source link, docs, or setup requirements do not match your workflow.

Risk explanation

No inherent user-impacting risk is flagged from the captured evidence.

Evidence links

Closest alternatives / related signals

agent-memorysymbolic-memorylong-term-memoryopenclawhermes-agenttencentwide-searchpersonamem