Overview
Shiliu focuses on cross-model, cross-agent Chinese long-term memory infrastructure, covering seven-path memory recall, temporal reasoning, conflict detection, expired-memory cleanup, context compression, and change provenance. The team led a unified memory API and SDK that closes the loop on write, recall, correction, and cross-agent sharing. On the LongMemEval-CN Chinese long-memory benchmark it passed 499/500 items in the first round (99.8% accuracy) and achieved 82.3% input-token compression in staged experiments. The product is now in deliverable Alpha, with public API, permission isolation, audit trails, and enterprise private deployment.
Connected profiles
More General-purpose agents products
Sources · 0Sources & provenance
English text is maintained separately from the original record. Consult the original source before citing a fact.
No source links are available for this profile.
Original source text
Names may use phonetic renderings when no English name is supplied. The original wording is retained below.
时流功能集中在跨模型、跨Agent的中文长期记忆基础设施研发，涵盖七路记忆召回、时间推理、信息冲突识别、过期记忆清理、上下文压缩及变更溯源等关键环节。团队主导完成了统一记忆API及SDK，实现记忆写入、召回、纠正和跨Agent共享闭环；在LongMemEval-CN中文长记忆测试中首轮通过499/500题，准确率达99.8%，并在阶段性实验中实现输入Token压缩82.3%。目前产品已进入可交付Alpha阶段，支持公网API、权限隔离、审计追踪及企业私有化部署。