The-AI-Republic/pi-dash

AI Agent Orchestration Platform

What it solves

Pi Dash is designed for "As Coding" (asynchronous vibe coding), a workflow where developers define high-level tasks and let AI coding agents handle the implementation in the background. It eliminates the need to manually monitor terminal outputs or "babysit" agent runs, allowing users to focus on scoping tasks and reviewing results.

How it works

Pi Dash uses a three-part architecture to orchestrate AI agents:

  1. Pi Dash Platform: A web-based hub for managing projects, defining tasks, organizing them into cycles and modules, and tracking progress via a dashboard.
  2. Pi Dash CLI & Runner Daemon: A local tool and background service that runs on the developer's machine. It polls the platform for assigned tasks, dispatches them to a configured AI agent, and reports the results back.
  3. AI Agent: The platform is agent-agnostic; users provide their own coding agent (with first-class support for Claude Code and Codex).

Who it’s for

Developers who want to automate the implementation phase of software development using AI agents without having to manually trigger and monitor every task.

Highlights

  • Agent-Agnostic: Supports various coding agents, including Claude Code and Codex.
  • Asynchronous Workflow: Shifts the focus from active monitoring to high-level task orchestration.
  • Self-Hostable: Can be deployed via Docker, Docker Compose, or an All-in-One image.
  • Integrated Skill: Offers an optional pi-dash-skill that allows agents to interact with Pi Dash issues directly.

Related

  • Project
    Grid2op/grid2op

    A modular platform for powergrid operation simulation, providing a Gymnasium-compatible environment for training reinforcement learning agents to manage power networks.

  • Project
    localai-org/apex-quant

    A MoE-aware mixed-precision quantization technique for llama.cpp that reduces MoE model size and increases speed while maintaining high accuracy.

  • Project
    modelstudioai/cli

    A command-line interface for Aliyun Model Studio that enables AI agents and developers to access multi-modal generation, model fine-tuning, and agent orchestration tools.

  • Project
    OnlyTerp/opengrok

    A configuration sidecar for Grok Bot that allows users to swap its default models for any OpenAI-compatible LLM with verified wire-mappings for optimal performance.