Pi Dash -- AI Agent Orchestration Platform

Live • Website • Documentation • Community • X

English • 简体中文 • 日本語 • Português (BR) • Español • 한국어

Pi Dash is an open-source AI agent orchestration platform built for As Coding (asynchronous vibe coding) — a workflow where you define what needs to be built, and coding agents handle the implementation in the background. Instead of babysitting agent runs and watching terminals scroll, Pi Dash lets you focus on the work that matters: scoping tasks, reviewing results, and shipping products.

Try it now at pidash.airepublic.com — no installation required.

Pi Dash is evolving every day. Your suggestions, ideas, and reported bugs help us immensely. Do not hesitate to open a GitHub discussion or raise an issue. We read everything and respond to most.

Pi Dash is composed of three major components:

The web-based orchestration hub where you manage projects, define tasks, and monitor agent progress. Create work items, organize them into cycles and modules, review agent output, and track analytics — all from a single dashboard. This is where you spend your time instead of watching terminals.

A local command-line tool and background daemon that runs on your development machine. The CLI connects to the Pi Dash platform, picks up assigned tasks, dispatches them to your configured AI agent, and reports results back. The runner daemon keeps this loop going continuously so you don't have to trigger each task manually.

Pi Dash is agent-agnostic — bring your own coding agent. Today the runner ships first-class support for Claude Code and Codex; the dispatch layer is designed so additional agents can be wired in without changing the orchestration model. You configure which agent the runner invokes; Pi Dash handles the rest.

Pi Dash Cloud is now live at pidash.airepublic.com — sign up to get started without running your own infrastructure. Prefer to self-host? Follow the steps below.

• Docker Engine installed and running
• Node.js version 22+ LTS version
• Python version 3.12+
• Postgres version v15+
• Valkey v7+ (or Redis 7+, drop-in compatible)
• Memory: Minimum 12 GB RAM recommended

  Running the project on a system with only 8 GB RAM may lead to setup failures or memory crashes (especially during Docker container build/start or dependency install). Use cloud environments like GitHub Codespaces or upgrade local RAM if possible.

1. Clone the repo

git clone https://github.com/The-AI-Republic/pi-dash.git [folder-name]
cd [folder-name]
chmod +x setup.sh

2. Run setup.sh

./setup.sh

setup.sh copies every .env.example to its .env counterpart (the repo root plus apps/web, apps/api, apps/space, apps/admin, apps/live), generates a unique Django SECRET_KEY and appends it to apps/api/.env, then runs pnpm install. For the default loopback-dev setup you do not need to edit any .env file manually — the .env.example defaults (localhost URLs, pi-dash database credentials, a local MinIO endpoint, etc.) work out of the box. Edit them only if you're binding to a non-default host/port or wiring in external services.

3. Start the containers

docker compose -f docker-compose-local.yml up

4. Start web apps:

pnpm dev

5. Open your browser to http://localhost:3001/god-mode/ and register yourself as instance admin
6. Open your browser to http://localhost:3000 and log in using the same credentials

For real self-hosted deployments (not local development), pick the path that matches how much you want to manage yourself:

• All-in-One Docker image — one container that bundles every Pi Dash service, managed by supervisord internally. Simplest path: a single docker run command. Best for demos, homelab setups, evaluation, and small teams. External Postgres / Redis / RabbitMQ / S3-compatible storage are still required.
• Docker Compose / Swarm self-hosting — the full microservices stack (6 service containers + database + queue + storage). More configuration, but gives you independent scaling and rolling updates per service. Recommended for anything beyond evaluation.
• Kubernetes / Helm — Helm chart publishing is planned but not yet shipped; see deployments/kubernetes/community/README.md.

Install the CLI on any machine where you want agents to pick up and execute tasks. Currently supported platforms:

• macOS — Apple Silicon (arm64)
• macOS — Intel (x86_64)
• Linux — arm64 and x86_64
• Windows — x86_64

On macOS and Linux, run the following command in your dev machine terminal:

curl --proto '=https' --tlsv1.2 -LsSf \
  https://github.com/The-AI-Republic/pi-dash/releases/latest/download/pidash-installer.sh | sh

On Windows, download and run the MSI installer:

https://github.com/The-AI-Republic/pi-dash/releases/latest/download/pidash-x86_64-pc-windows-msvc.msi

To pin to a specific version (or install a prerelease), swap latest for the tag, e.g. .../releases/download/pidash-v0.1.4/pidash-installer.sh. Prereleases are excluded from /latest/, so the one-liners above always serve the last stable release. Full pinning recipes — wrapper, bare installer, Windows variants — are in runner/README.md.

Then authenticate the machine and register a runner. The standard flow is two commands:

# 1. Browser-based device-code login (like `gh auth login` / `stripe login`).
#    Stores a CLI token at ~/.config/pidash/config.toml.
pidash login --url https://your-pidash-instance.com

# 2. Register this host as a runner. Uses the token from step 1 — no
#    enrollment-token paste needed. On the first runner, installs the OS
#    service (systemd user unit on Linux, launchd agent on macOS, or a
#    per-user scheduled task on Windows) and starts the daemon.
pidash runner add --project <project-id>

pidash login (an alias of pidash auth login) prompts to add a runner inline when no runner exists yet, so a fresh dev laptop can be onboarded with a single command. Add more runners later with pidash runner add --project <other-project-id>.

The runner daemon runs in the background, polls for assigned tasks, dispatches them to your AI agent, and reports results back to the platform.

Useful commands:

See pidash --help for all available commands.

Pi Dash does not ship an AI agent — you bring your own. Ensure your chosen agent CLI is installed and accessible on the machine running the Pi Dash CLI. The runner currently supports two agent kinds out of the box:

• Claude Code — install claude and make sure claude --version works.
• Codex — install codex and make sure codex --version works.

pidash doctor verifies the configured agent is on PATH and the cloud is reachable before you go live.

pi-dash-skill packages a portable agent skill so Claude Code or Codex can create, list, move, and inspect Pi Dash issues directly from a coding session via the pidash CLI.

npx @airepublic/pidash-skill-installer           # installs to Claude Code, Codex, or both

The installer prompts for a target (default: all) and fetches the skill from GitHub on demand — no clone required. Pass --all, --claude-code, or --codex to skip the prompt.

Codex users can also install via the built-in $skill-installer from inside a Codex session. See the pi-dash-skill README for env-var overrides (CLAUDE_HOME / CODEX_HOME), the clone-based install path, and other alternatives.

Explore the Pi Dash documentation to learn about features, setup, and usage.

Join the conversation on GitHub Discussions, follow @ai_republic on X, or visit airepublic.com for updates. We follow a Code of conduct in all our community channels.

Feel free to ask questions, report bugs, participate in discussions, share ideas, request features, or showcase your projects. We'd love to hear from you!

If you discover a security vulnerability in Pi Dash, please report it responsibly instead of opening a public issue. See SECURITY.md for more info.

To disclose any security issues, please email us at privacy_security@airepublic.com.

There are many ways you can contribute to Pi Dash:

• Report bugs or submit feature requests.
• Review the documentation and submit pull requests to improve it—whether it's fixing typos or adding new content.
• Show your support by upvoting popular feature requests.

Please read CONTRIBUTING.md for details on the process for submitting pull requests.

Pi Dash is built on top of Plane, an open-source project management tool. We are grateful to the Plane team and its contributors for laying the foundation that made Pi Dash possible.

This project is licensed under the GNU Affero General Public License v3.0.