Runs a self-hosted personal assistant that answers on DingTalk, Lark, WeChat, Discord, Telegram and more, with layered memory, skills, cron jobs, a sandbox and optional local models.
Skills
Multi-Channel Chat
Serves one assistant instance across DingTalk, Lark, WeChat, Discord, Telegram, iMessage and QQ, plus a web console, TUI and desktop app.
Layered Memory
Keeps live working context, full verbatim history and a self-evolving Markdown knowledge base built on ReMe across conversations.
Scheduled Tasks
Sets up recurring jobs such as news digests, report generation and multi-channel broadcasts through QwenPaw's cron configuration.
Local Models
Runs QwenPaw-Flash 2B, 4B or 9B models in the built-in llama.cpp runtime, or Ollama and LM Studio, with no API key needed.
Related Agents
HumanLayer
Gates autonomous agent tool calls behind human approval in Slack or email, so high-risk actions require a person to sig…
CowAgent
Runs a self-hosted personal agent that plans multi-step tasks, calls built-in and MCP tools, installs skills, and answe…
Spacebot
Runs a multi-user agent harness for Discord, Slack, Telegram, and Twitch with shared graph memory, tasks, cron jobs, an…
OpenClaw
Runs a self-hosted, local-first personal AI assistant across WhatsApp, Telegram, Slack, and Discord. Ships a CLI gatewa…