helix

♾️ Private Agent Fleet with Spec Coding. Each agent gets their own GPU-accelerated desktop. Run Claude, Codex, Gemini and open models on a full private AI Stack

What is helix?

♾️ Private Agent Fleet with Spec Coding. Each agent gets their own GPU-accelerated desktop. Run Claude, Codex, Gemini and open models on a full private AI Stack ♾️

Overview of helix

What it is

Helix is a private agent fleet platform that runs coding agents on your own infrastructure. Each agent gets its own GPU-accelerated desktop with a browser, terminal, filesystem, and GUI apps. Work is organized on a spec-driven Kanban board, and agents open pull requests when tasks are complete. Helix supports multiple agent harnesses (Claude Code, OpenAI Codex, Gemini CLI, Qwen Code, Goose, Zed Agent, and any ACP-compatible agent) and multiple LLM providers, including self-hosted models via vLLM. It is self-hostable end to end, including air-gapped deployments.

Key features

  • Full desktop per agent (GPU-accelerated, streaming) with browser, terminal, filesystem, and GUI apps - Fleet visibility dashboard to watch any agent's live screen and jump in with pair programming - Multiplayer: agent environments are shared across teammates, with persistent chat history and state - High-density isolation: many fully isolated agent desktops on a single machine with deduplicated filesystem and per-agent credential/network isolation - No lock-in: swap agent harnesses per task and point Helix at any LLM provider, including self-hosted models - Spec-driven Kanban board with columns: Backlog → Planning → Spec Review → In Progress → Pull Request → Merged - Parallel task execution with per-column WIP limits

Use cases

  • Running a fleet of coding agents on your own servers instead of on individual developer laptops - Collaborative agent-driven development where a team dispatches, steers, and reviews many agents from a shared board - Spec-driven development: describe an outcome, an agent plans it, you approve, and it implements and opens a PR - Air-gapped or private deployments where all infrastructure and models are self-hosted - Using cheaper local models for initial planning and escalating to frontier models only when needed

Getting started

Install on Docker: bash curl -sL -O https://get.helixml.tech/install.sh chmod +x install.sh sudo ./install.sh The dashboard will be available on http://localhost:8080 by default. For Kubernetes, use the provided Helm charts (Control Plane and Runner). For local development: bash git clone https://github.com/helixml/helix.git cd helix docker-compose up -d postgres cd api go run . serve # In another terminal: cd frontend npm install npm run dev

License and pricing

The license is not stated in the repository metadata (NOASSERTION). The project is listed as open source.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/helixmlhelix.svg)](https://yunodb.com/tools/helixmlhelix?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/helixmlhelix?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/helixmlhelix.svg" alt="Featured on YunoDB" height="20"></a>