Build a private assistant that summarises your email and calendar each morning and delivers a digest to your Telegram.
Set up a multi-agent pipeline where one agent researches topics overnight and another drafts a written report by morning.
Write a custom skill that watches a folder for new PDF files, reads them, and sends you a summary on Discord.
Connect the same assistant to WeChat, DingTalk, and Feishu so you can chat with it from any app without duplicating config.
Requires Python 3.10, 3.13, a self-managed server or local machine, and API credentials for each chat channel you want to connect.
QwenPaw is a personal AI assistant that you can install on your own computer or on a cloud server you control, instead of using a chat product hosted by someone else. The pitch in the README is "works for you, grows with you", the assistant is meant to remember things about you, learn from the conversations you have with it, and act on your behalf over time. Because you deploy it yourself, your data and memory stay on your own machine or on a server you choose, with no third-party hosting and no data upload to an outside vendor. The assistant is built around the idea of skills, which are the individual capabilities that decide what it can actually do. It ships with built-in skills like scheduling, reading and processing PDF and office documents, and producing news digests, and you can add your own custom skills that get loaded automatically. It also supports multi-agent collaboration: you can create several agents with different roles and let them talk to each other to handle more complex tasks. A multi-layer security system (a tool guard, file access control, and skill security scanning) is included to keep things from going wrong when an agent is acting on your behalf. A key practical feature is that QwenPaw can connect to many chat channels, DingTalk, Feishu, WeChat, Discord, Telegram and others, so you can talk to the same assistant from whichever app you prefer. Example uses listed in the README include daily social-media digests, email and calendar summarisation, prototype generation while you sleep, tracking tech and AI news, and organising or searching local files. Someone would use this if they want a private, programmable AI assistant they fully control. It is a Python project (3.10 up to but not including 3.14), licensed under Apache 2.0, was recently renamed from CoPaw to QwenPaw, and is being actively developed. The full README is longer than what was provided.
← agentscope-ai on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.