<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> Lingtai provides a local-first framework for building persistent AI organizations, moving beyond ephemeral agents. It targets developers and power users seeking robust, inspectable, and recoverable AI collaborators that maintain state, memory, and communication across sessions, enabling complex, long-running AI workflows.

How It Works

Lingtai establishes AI organizations through resident agents with dedicated filesystem homes (.lingtai/), durable memory, and inter-agent communication via mailboxes. Its core innovation is acting as an organizational layer, managing agent roles, memory, supervision, and recovery around execution tools like CLIs or coding agents. A "soul flow" allows idle agents to reflect proactively, while agents can spawn persistent "avatars" or short-lived "daemons" to scale tasks. Communication integrates seamlessly with a TUI, Telegram, Feishu, WeChat, WhatsApp, and email, all channeling into the same persistent organization.

Quick Start & Requirements

• Install: Recommended via Homebrew: brew install lingtai-ai/lingtai/lingtai-tui.
• Prerequisites: Homebrew, Go (for source builds), Node.js/npm (for portal). The TUI manages its own Python runtime.
• Setup: Initial run guides model/preset configuration and initializes the .lingtai/ project directory.
• Links: Website, Releases.

Highlighted Details

• Long-lived, local-first agents with persistent filesystem homes, mailboxes, and durable knowledge bases.
• "Soul flow" enables proactive reflection, self-correction, and emergent next steps.
• Dynamic multi-agent network formation (avatars/daemons) visualized via lingtai-portal.
• Unified communication across TUI, Telegram, Feishu, WeChat, WhatsApp, and email, all interacting with the same agent organization.
• Leverages coding agents (e.g., Claude Code, Codex) as execution hands while Lingtai manages the overarching plan, memory, and coordination.

Maintenance & Community

Lingtai is a dual-repository project: Lingtai-AI/lingtai for the TUI and portal, and Lingtai-AI/lingtai-kernel for the Python agent runtime. Community engagement is fostered through Discord, GitHub Issues, and Discussions. The project appears actively maintained, with detailed contribution guidelines and a clear architecture.

Licensing & Compatibility

The project is licensed under the permissive Apache-2.0 license, making it suitable for commercial use and integration within closed-source applications without copyleft restrictions.

Limitations & Caveats

While feature-rich, the README does not explicitly detail limitations. Managing sophisticated multi-agent networks and their emergent behaviors may present a learning curve. Source builds require development tools like Go and Node.js, potentially adding setup friction. The practical effectiveness and controllability of the "soul flow" proactive reflection mechanism would require evaluation.