Skip to content

🤖 Jarvis

Personal AI Infrastructure

An open-source, self-hosted, Iron Man-style personal AI assistant living across all your devices.

What is Jarvis

Jarvis is a personal AI infrastructure that brings a unique, persistent and context-aware assistant to every device you own — inspired by Iron Man's J.A.R.V.I.S.

It's not just a chatbot: it's a distributed network of specialised agents living on laptops, smartphones, smartwatches, smart glasses, VR headsets, holographic displays and medical wearables, all coordinated by a single private AI identity.

Key features

  • Web search & scraping


    Agents that fetch, summarise and cite web sources in real time. Stack: Crawl4AI, Firecrawl, Jina Reader, ScrapeGraphAI.

  • Cross-device voice


    Wake-word, speech-to-text and natural dialogue on any device. Stack: Porcupine, openWakeWord, faster-whisper, Piper TTS.

  • Augmented reality


    Informational overlays on Meta Ray-Ban, XREAL, Brilliant Frame and OpenXR headsets. Stack: Frame SDK, MentraOS, XREAL SDK.

  • Virtual reality


    Immersive interface on Quest, Valve Index and any OpenXR-compatible headset. Stack: OpenXR, Monado.

  • Holographic displays


    Integration with Looking Glass, Voxon and volumetric displays. Stack: Looking Glass Core SDK, Voxon SDK.

  • Medical wearables


    Apple Watch, Wear OS, Garmin, Whoop, Oura — biometric data via HealthKit, Health Connect and FHIR.

  • Persistent memory


    The conversation follows you across devices thanks to mem0, Qdrant, Zep.

  • Privacy by design


    Self-hosted, data under your control, local or cloud models at your choice.

Quick start

git clone https://github.com/fedcal/open-jarvis.git
cd open-jarvis
cp .env.example .env
docker compose up -d

Getting started Architecture

Project status

Work in progress

We are defining the architecture and the first MVPs. We are looking for contributors — see the Contributing page.

Philosophy

The value is not the chatbot. The real value is a persistent AI identity that follows the user across every device.