taskuary

Taskuary is an open-source, local-first AI task hub designed to automate job workflows. It aggregates incoming messages from email, Teams, Slack, issue trackers, alerts, and reports into a single timeline, categorizing them as Urgent, On you, For later, Advisor ideas, and FYI. An AI Assistant walks users through tasks, and AI agents (connected via coding CLIs like Claude Code, Codex, Qwen Code, OpenCode, Kimi Code, Gemini, Cursor, Copilot, Muse Code, or others) can perform work such as analysis and drafting replies. Users retain final approval before anything is sent or shipped. Key features include: - **Unified timeline**: All work lands in one place, sorted by priority and need. - **AI triage and assistance**: The Assistant explains tasks, suggests next actions, and answers questions using context from messages, threads, and past decisions. - **Agent integration**: Connect coding CLIs to run agents that work inside the chat, with session tracking and review. - **Shared Hub**: Agents share discoveries, decisions, and warnings by topic, enabling collaborative learning. - **Live handoffs**: Agents leave notes for each other on a Board, showing progress and blockers. - **Adaptive memory**: Learns from user corrections and edits, building lessons in LEARNED.md to personalize future behavior. - **Privacy and security**: Runs locally; credentials are redacted from prompts before leaving the machine, and replies containing placeholders are refused. Installation options include a single-file Windows executable, Python (pip install taskuary), or Docker. A demo mode is available for trying the interface without connecting external systems. The project is in early development (v0.3.7.2) and licensed under MIT.