← Back
Open Source📊 data
TencentDB Agent Memory
Fully local long-term memory for AI agents via a 4-tier progressive pipeline with zero external API dependencies.
github.com →📊Categorydata
🆓PricingOpen Source
About
TencentDB Agent Memory combines symbolic short-term memory with layered long-term memory to help AI agents retain task context and reuse past experience across sessions. Symbolic memory condenses heavy tool logs into compact Mermaid symbols to cut token usage, while layered memory distills conversations into structured personas and scenes. When integrated with OpenClaw, it reduces tokens by up to 61% and improves task success and persona accuracy, making it useful for developers building long-horizon AI agents.
Pricing
🌐
