Your personal AI. Self-hosted. Zero cloud.

See it in action

Why Hephaistos

Self-Hosted Privacy

Your data never leaves your network. All inference runs locally on your hardware. Zero telemetry, zero tracking.

Multi-Platform

Native apps for iOS, Android, Windows, macOS and Linux. One server, every device. Seamless experience.

Open Source

Transparent codebase. Audit, extend, contribute. Built with standard tools you already know and trust.

Agents & Tools

Create AI agents with custom tools. Web search, code execution, calculators. Chain agents for complex workflows.

Voice Chat

Talk to your AI with speech-to-text and text-to-speech. Hands-free conversations, natural interaction.

End-to-End Encryption

WireGuard VPN tunnel between devices. TLS encryption on all communications. Your conversations stay yours.

How it works

1

Install

Run the installer on any machine with a GPU. One script sets up the LLM, the API, and the database.

2

Connect

Open the app on your phone or computer. Scan the network or enter your server address. WireGuard VPN included.

3

Chat

Talk to your AI. It remembers context, runs tools, and stays private. All inference happens on your hardware.

Simple pricing

Self-host for free. Or let us handle the hardware.

Self-Hosted

Free

0€ / forever

Bring your own hardware

  • Full app access
  • All features included
  • Unlimited conversations
  • Community support
  • Open source
Get Started

Install in one command

Open a terminal and paste. That's it.

Terminal.app
curl -fsSL https://hephaistos.olympe.tech/install.sh | bash

Requires 16GB+ RAM. Installs Ollama + Gemma 4 + Hephaistos backend.
Detects your GPU automatically (NVIDIA, AMD, Apple Silicon).

Support the project

Hephaistos is free and open-source. If you find it useful, consider buying us a coffee.

Buy us a coffee

Built with

Flutter
Python
FastAPI
Ollama
Supabase
WireGuard