TencentDB Agent Memory
🔥 This release made it to our blog
TencentDB Agent Memory: Layered Long-Term Memory for AI Agents With 61% Token Reduction
Tencent has open-sourced TencentDB Agent Memory, a fully local long-term memory system for AI agents built on a 4-tier progressive pipeline combining symbolic short-term memory and layered long-term personalization. In benchmarks, it cuts token usage by up to 61.38% and improves task success rates by 51.52%, with zero external API dependencies.
### TL;DR
TencentDB Agent Memory is a fully local, open-source long-term memory system for AI agents built by Tencent. Using a 4-tier progressive pipeline with symbolic short-term memory (Mermaid canvas) and layered long-term memory (L0→L3 persona hierarchy), it cuts token usage by 61% and improves agent task success by 51% when integrated with OpenClaw — with zero external API dependencies.
Key Insights & Metrics
Key Features
- Symbolic short-term memory — offloads verbose tool logs to external files and condenses in-context state into compact Mermaid symbol graphs with node_id tracing; cuts WideSearch token usage by 61.38% and improves pass rate from 33% to 50% (relative +51.52%)
- 4-tier long-term memory pipeline — L0 raw conversation → L1 atomic facts → L2 scenario blocks → L3 user persona; layered storage with full drill-down traceability (no irreversible compression); raises PersonaMem accuracy from 48% to 76%
- Fully local, zero external API dependencies — dual-layer storage (database for facts/logs + Markdown for personas/scenes); integrates as an OpenClaw plugin and Hermes skillpack; SWE-bench improvement +9.93% with 33% token savings
→ Related Releases
MiroThinker
MiroThinker is an open-source search agent model developed by MiroMindAI, designed for tool-augmented reasoning and real-world information seeking. It aims to match the deep research capabilities of leading AI models like OpenAI's Deep Research and Google's Gemini Deep Research.
Letta Code SDK
The Letta Code SDK is a software development kit that enables developers to build deeply personalized agents with persistent memory that learn over time. It serves as the interface to Letta Code, facilitating the creation of stateful agents capable of continuous learning and improvement.
OB-1
OB-1 is a self-improving coding agent developed by OpenBlock Labs, designed to autonomously handle the full development lifecycle, from project management to pull requests. It integrates seamlessly into existing workflows, enhancing productivity and code quality.
AIO Sandbox
AIO Sandbox is an integrated environment designed for AI agents, combining a browser, terminal, filesystem, VSCode, Jupyter, and MCP Server into a single Docker container. This unified setup allows seamless development and execution of AI agents without the need for multiple services.
Discussion
Sign in to leave a review
Reviews
No reviews yet. Be the first to review!