
ppl - Personal CRM for AI Agents
io.github.dbhurleyv1.2.0更新于 Oct 3, 2026
Personal CRM operated by AI agents. Read relationship context before acting, write back after.
概览
一个托管型个人 CRM,让 AI 助手读取人际关系背景,并记录联系人、笔记、任务和互动。
- 功能
- ppl 是一个专为 AI 助手操作而设计的托管型个人 CRM,助手充当智能层,ppl 充当记忆与数据层。它提供约 80 个工具,涵盖每日简报、语义搜索、人物摘要、建议操作和互动记录。它还提供对联系人、笔记、任务、提醒、通话、礼物、日记条目、地址、关系、地点、债务、文档和生活事件的完整增删改查操作。
- 适用场景
- 当你希望助手维护个人关系记录时值得添加:在起草消息或安排通话前查询某人的背景信息,事后再写回记录。适合需要管理大量联系人并希望获得简报、摘要和跟进建议的人。
- 运行要求
- 通过 Streamable HTTP 连接 ppl.gift 的远程 MCP 端点。认证使用 Bearer 令牌,即从 ppl.gift 设置中获取的每用户 API 令牌。无需本地运行时或安装包。
安装
在 SourceWeft 中
- 打开 控制台中的 ppl - Personal CRM for AI Agents,将其添加到工作区。
- 为需要使用其工具的对话启用该服务。
Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。
其他 MCP 客户端
把它添加到你客户端的 mcpServers 配置中。
{
"mcpServers": {
"ppl": {
"type": "http",
"url": "https://ppl.gift/mcp"
}
}
}README
ppl
The personal CRM for AI agents.
ppl is a personal CRM designed to be operated by AI agents. Your AI is the intelligence layer. ppl is the memory and data layer. The agent reads relationship context before acting, and writes back after.
The hosted service runs at ppl.gift. This repo holds the public agent interface: the MCP server contract, integration docs, and examples. The application itself is private.
Connect your agent
ppl exposes a Model Context Protocol server at:
Transport: Streamable HTTP. Auth: Bearer token (per-user API token from ppl.gift settings).
80 tools covering: daily briefings, semantic search, person digests, suggested actions, interaction logging, and full CRUD for contacts, notes, tasks, reminders, calls, gifts, journal entries, addresses, relationships, places, debts, documents, and life events.
See docs/agents.md for the full integration guide.
Examples
- examples/python — minimal MCP client in Python
- examples/node — minimal MCP client in Node.js
Registry
server.json is the MCP registry manifest for this server.
License
MIT. See LICENSE.
来源:README.md,提交 de5c33d
工具
0版本历史
1- v1.2.0最新Oct 3, 2026