Knowledge2026年5月12日·1 分钟阅读

yantrikdb-server — Memory DB with Contradiction Checks

YantrikDB adds persistent agent memory with contradiction checks. Use `yantrikdb-mcp` for Claude Code/Cursor, or scale with a server cluster.

Agent 就绪

这个资产可以被 Agent 直接读取和安装

TokRepo 同时提供 CLI 命令、metadata JSON、安装计划和原始内容链接,方便 Agent 判断适配度、风险和下一步动作。

Needs Confirmation · 62/100策略:需确认
目标
Claude Code, Cursor, Windsurf
类型
Mcp
安装
Pip
信任
信任等级:Established
入口
pip install yantrikdb-mcp && yantrikdb-mcp --help
CLI 安装命令
npx tokrepo install 84ec90f4-1f41-5595-ba10-fedfb0b8bc40 --target codex

简介

yantrikdb-server 侧重部署与集群化:作为认知型记忆数据库,它能帮 Agent 保留决策、提示“变陈旧”的重要记忆,并自动标记矛盾点。

最适合: 需要可治理的持久记忆(陈旧提示 + 矛盾检测)的 Agent

适配: 默认本地 SQLite 记忆;可选集群(server URL + token);MCP 客户端(Claude Code/Cursor/Windsurf)

配置时间: 10–25 分钟

关键事实(已验证)

  • README 快速开始:pip install yantrikdb-mcp + 一段配置即可启用 MCP。
  • README 表示重启后会暴露 15 个记忆工具。
  • GitHub:142 stars · 9 forks;最近更新 2026-05-12(GitHub API 验证)。

正文

为了让记忆“有用而不吵闹”:

  • 先用本地 SQLite 验证 Agent 会记住什么,再在 schema 与访问策略明确后切到共享集群。
  • 把矛盾标记当作“需要澄清”的提示,而不是自动裁决真假。
  • 定期回顾写入策略:优先记稳定偏好/决策/约束,不要无脑记录全量对话。

README 原文节选(verbatim)

YantrikDB

Correction notice (2026-04-19): Phase 3 benchmark writeups committed earlier today (docs/phase3a/, docs/phase3b/, docs/phase3c/, docs/phase3d/) used a Python simulator for the "structured memory" condition — NOT the actual yantrikdb engine. Rerun with real yantrikdb is in docs/phase3e/. Full audit trail in CORRECTIONS.md. Full corrected findings post coming 2026-04-20.

A memory database that forgets, consolidates, and detects contradictions.

Vector databases store memories. They don't manage them. After 10,000 memories, recall quality degrades because there's no consolidation, no forgetting, no conflict resolution. Your AI agent just gets noisier.

YantrikDB is different. It's a cognitive memory engine — embed it, run it as a server, or connect via MCP. It thinks about what it stores.

Shortest path to try it: MCP setup for Claude Code / Cursor / Windsurf → (one pip install, one config block).

The bigger picture: YantrikDB is the memory layer being built on the road to YantrikOS — an AI-native operating system where agents are first-class primitives, not apps on top. Memory was the bottleneck, so we're shipping it first.

Crates.io PyPI Docker DOI License: AGPL-3.0

📄 New paper (May 2026): Skill as Memory, Not Document — measuring three failure modes when agent skill catalogs scale to 5,000 skills. Code + raw CSVs reproducible in benchmarks/skill_recall/.

YantrikDB demo: storing three facts, recalling them, then think() flagging a contradiction between two memories

99.9% token savings vs file-based memory

Memories File-Based (CLAUDE.md) YantrikDB Token Savings Recall Precision
100 1,770 tokens 69 tokens 96% 66%
500 9,807 tokens 72 tokens 99.3% 77%

FAQ

MCP server 是单独的组件吗? 答:README 指向 yantrikdb-mcp 作为 Claude Code/Cursor/Windsurf 的推荐起步方案。

默认记忆存在哪里? 答:README 表示默认使用 ~/.yantrikdb/memory.db 的本地 SQLite 文件。

怎么共享记忆? 答:README 展示通过 YANTRIKDB_SERVER_URLYANTRIKDB_TOKEN 配置集群模式。

🙏

来源与感谢

Source: https://github.com/yantrikos/yantrikdb-server > License: AGPL-3.0 > GitHub stars: 142 · forks: 9

讨论

登录后参与讨论。
还没有评论,来写第一条吧。

相关资产