{"$schema":"https://wellknown.network/schemas/agent-record-v1.json","schemaVersion":"1","id":"ag_scxmecys4asv","handle":"kb-mcp-lite","url":"https://wellknown.network/agents/kb-mcp-lite","links":{"self":"https://wellknown.network/agents/kb-mcp-lite/record.json","html":"https://wellknown.network/agents/kb-mcp-lite","markdown":"https://wellknown.network/agents/kb-mcp-lite/record.md","api":"https://wellknown.network/api/v1/agents/kb-mcp-lite","status":"https://wellknown.network/api/v1/agents/kb-mcp-lite/status","claim":"https://wellknown.network/agents/kb-mcp-lite/claim","claimApi":"https://wellknown.network/api/v1/claims","claimDescriptor":"https://wellknown.network/agents/kb-mcp-lite/claim.json","badge":"https://wellknown.network/agents/kb-mcp-lite/badge.svg","openapi":"https://wellknown.network/openapi.json","history":"https://wellknown.network/api/v1/agents/kb-mcp-lite/history","tools":"https://wellknown.network/api/v1/agents/kb-mcp-lite/tools"},"ard":{"identifier":"urn:air::server:kb-mcp-lite","type":"application/mcp-server-card+json"},"kind":"mcp_server","declared":{"name":"kb-mcp-lite","summary":"Lightweight agent-native knowledge base. SQLite + FTS5 + vec0 + MCP server.","description":"<div align=\"center\">\n\n# kb-mcp-lite\n**面向AI代理的轻量级本地知识库 · 团队协作友好**\n\n`pip install kb-mcp-lite` — 让任何AI编程助手都拥有结构化、可查询、可同步的团队\"第二大脑\"\n\n[![PyPI version](https://img.shields.io/badge/pypi-v0.9.2-blue)](https://pypi.org/project/kb-mcp-lite/)\n[![Python](https://img.shields.io/badge/python-≥3.10-blue)](https://www.python.org/)\n[![License: MIT](https://img.shields.io/badge/license-MIT-green)](./LICENSE)\n[![MCP](https://img.shields.io/badge/MCP-兼容-purple)](https://modelcontextprotocol.io/)\n[![Status: 测试版](https://img.shields.io/badge/status-beta-green)](#状态说明)\n\n</div>\n\n---\n\n## 🎯 解决什么问题\n当前的知识管理工具存在明显的断层：\n- 面向人类的知识库（Notion/Obsidian）：需要手动维护，AI无法直接高效访问\n- 向量数据库：需要复杂配置，没有结构化Schema，团队协作困难\n- 团队文档散落在各个地方：代码注释、Wiki、PR描述、飞书文档，AI找不到也用不了\n\n`kb-mcp-lite` 专门填补这个空白：**专为AI代理设计，同时兼顾人类编辑和团队协作**，让AI编程助手可以直接调用团队沉淀的所有技术知识。\n\n| 对比维度 | Notion/Obsidian | Chroma/LanceDB等向量库 | **kb-mcp-lite** |\n|---|---|---|---|\n| 服务对象 | 人类 | 模型嵌入 | **AI代理 + 人类开发者** |\n| 访问方式 | Web UI | SDK调用 | **MCP标准协议（AI原生） + CLI + Web管理后台** |\n| 结构规范 | 自由格式 | 无结构 | **强Schema（标准化文档类型）** |\n| 存储方案 | 云服务/私有部署 | 本地文件 | **SQLite + FTS5 单文件存储** |\n| 团队协作 | 在线协作 | 无协作能力 | **Git原生同步，纯文本版本管理** |\n| 部署成本 | 账号/服务器 | 安装配置 | **`pip install` 开箱即用，零配置** |\n| 数据隐私 | 云服务存储 | 本地 | **完全本地存储，无云服务，无遥测** |\n\n---\n\n## ✨ 核心能力详解\n### 1. 📐 强Schema标准化与结构化Metadata\n内置9种开箱即用的文档类型，覆盖技术团队90%的知识沉淀场景，所有类型支持结构化Typed Metadata并可通过Python子类扩展：\n| 文档类型 | ID前缀 | 用途说明 | 结构化Metadata亮点 |\n|---|---|---|---|\n| `project` | `proj` | 项目/仓库说明文档 | 记录技术栈、负责人、状态、部署流程 |\n| `decision` | `dec` | 架构决策记录(ADR) | 结构化记录 `status`(accepted/superseded)、决策人、替代方案 |\n| `lesson` | `lesson` | 经验教训/踩坑记录 | 结构化记录 `severity`、根因归类、避坑准则 |\n| `glossary` | `glossary` | 术语表 | 统一业务/技术专有名词标准定义 |\n| `person` | `person` | 人员档案 | 团队成员技术栈、模块分工与联系方式 |\n| `faq` | `faq` | 常见问题 | 沉淀高频问答，消除重复咨询 |\n| `api` | `api` | API接口文档 | 结构化记录 HTTP Method、Path、鉴权与限流要求 |\n| `runbook` | `runbook` | 运维手册(SOP) | 标准化运维及应急处置步骤 |\n| `release` | `release` | 发布日志 | 版本号、变更清单、影响范围与回滚步骤 |\n\n**优势**：\n- **真实结构化元数据**：支持 JSON 字段与 Frontmatter 无损互转，AI 可精确抽取与检索事实；\n- 自动生成稳定I…","publisher":{"name":"HelloTomBruce (Zhang Bei)","url":null},"homepage":"https://github.com/HelloTomBruce/kb-mcp-lite","repository":"https://github.com/HelloTomBruce/kb-mcp-lite/releases","version":"0.9.2","license":"MIT","protocols":["mcp"],"tags":["knowledge-base","mcp","agent","sqlite","fts5","vector-search","semantic-search","lightweight"],"pricing":null,"endpoints":[{"url":"pypi:kb-mcp-lite","type":"package_pypi","auth":null,"probeable":false}],"skills":null,"tools":null,"extra":null,"attribution":{"kind":"pypi","name":"pypi","license":"pypi","repoUrl":"pypi","summary":"pypi","version":"pypi","description":"pypi","homepageUrl":"pypi","publisherName":"pypi"}},"derived":{"capabilities":[{"slug":"data.database","name":"Databases","confidence":1,"provenance":"declared"},{"slug":"data.vector-search","name":"Vector Search","confidence":1,"provenance":"declared"},{"slug":"productivity.notes","name":"Notes & Docs","confidence":1,"provenance":"derived"}],"categories":["data","productivity"],"language":"en"},"observed":{"status":"unknown","statusReason":"Distributed as a package to run locally; no network endpoint to check.","lastOkAt":null,"lastProbedAt":null,"statusComputedAt":null,"reliability30d":null,"latestObservations":[],"tools":null,"package":{"name":"kb-mcp-lite","registry":"pypi","observedAt":"2026-11-09T12:23:40.459Z","publishedAt":"2026-11-02T06:25:37.149444Z","latestVersion":"0.9.2"},"toolSurface":null,"endpointFacts":[]},"verification":{"claimed":false,"claimedAt":null,"proofs":[]},"provenance":{"sources":[{"source":"pypi","key":"kb-mcp-lite","url":"https://pypi.org/project/kb-mcp-lite/","firstSeenAt":"2026-09-09T22:23:09.609Z","fetchedAt":"2026-11-09T12:22:07.370Z","normalizedAt":"2026-11-09T12:22:07.370Z"}]},"firstSeenAt":"2026-09-09T22:23:09.609Z","updatedAt":"2026-11-09T12:23:40.459Z"}