Tencent Open-Sources Agent Memory System, Reduces Token Usage Up to 61%

According to Beating, Tencent Cloud recently open-sourced TencentDB Agent Memory, a local-first memory engine for AI agents. The system reduces token consumption by 61% in complex workflows—decreasing usage from 221.31M to 85.64M in WideSearch tasks—while improving task completion rates by 51.52%. The engine uses a layered memory architecture separating long-term memory (conversations, atomic facts, scenario chunks, and user profiles) from short-term task memory, with logs externalized and tasks visualized via Mermaid diagrams for efficient retrieval.
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments