Self-host J.A.R.V.I.S
Iron-Man-style voice assistant with holographic HUD
- Stars
- 198
- CPU
- Apple Silicon (tested) / Linux
- Storage
- ~460 MB for Whisper small.en model
- Ports
- 8765 (Plain WebSocket for push-to-talk client), 443 (HTTPS/WSS for browser HUD (microphone requires TLS)), 8766 (Additional HTTPS/WSS port for browser HUD), 9443 (TLS reverse proxy for Hermes dashboard iframe)
- Requires
- ElevenLabs API, Hermes Agent
What is J.A.R.V.I.S?
This open-source project lets you run a voice-controlled AI assistant on your own hardware. It connects a local Whisper speech-to-text engine with ElevenLabs text-to-speech and wraps everything in a browser-based holographic heads-up display. The HUD shows live transcription, visualizes what the underlying Hermes Agent is doing, lets you stop or approve risky actions, and can summon video panels on command. A FastAPI server handles the audio streaming, proxies the agent's API, and serves the interface over TLS. You need a machine running macOS or Linux with Python 3.11, an ElevenLabs API key, and a separate Hermes Agent instance. The code is MIT-licensed and still in early development.
Should you self-host J.A.R.V.I.S?
J.A.R.V.I.S scores 38/100 on selfhost.directory (how scoring works) and ranks #10 of 11 in Voice Assistants. It has 198 GitHub stars.
There was no stable release in the past 12 months, so check the repository's recent activity before you rely on it.
Setup is advanced: plan for several services and some ongoing maintenance. It is MIT-licensed and written mostly in Python.
Its closest alternative on selfhost.directory is dograh (70/100, 5 972 stars). It scores higher overall, so compare the two before you pick. See J.A.R.V.I.S vs dograh side by side.
Features
- Agent Control
- Cinematic Boot
- Embedded Dashboards
- Holographic HUD
- Interrupt-Aware Barge-In
- Local Speech-to-Text
- Machine Stats Panel
- Optional GPU Acceleration
- Persistent conversation memory
- Privacy Filter
- Streaming Text-to-Speech
- Typed Chat
- Usage Tracking
- Voice assistant
J.A.R.V.I.S README
More in Voice Assistants
JasperSelf-hosted, open-source voice assistant you run on a Raspberry Pi, with pluggable speech recognition, text-to-speech, and Python modules for weather, news and Spotify.
LeonAn open-source personal assistant you run self-hosted on your own machine, with voice chat, offline STT/TTS, modular skills, layered memory, and agentic tool execution.
minutesFree, open-source app that records and transcribes meetings locally into Markdown, then lets AI assistants like Claude Code search the history via MCP.
mirrormateSelf-hosted voice-first AI companion for smart mirrors. Wake-word activation, RAG memory, expressive avatar, local inference with Ollama and VOICEVOX.
MoltisSelf-hosted, open-source AI agent server in a single Rust binary: multi-provider LLMs, voice, memory, MCP tools, and Telegram/Discord/Slack channels.
openclaw-voiceSelf-hosted browser voice interface for AI assistants. Runs Whisper locally, uses ElevenLabs or local TTS, connects to OpenAI, Claude, or custom agents.
sapphireSapphire Blue AI is a self-hosted, open-source AI companion with voice, memory, personas, and plugins. Runs locally with optional cloud LLMs.
SayItSayIt is an open-source Windows voice typing app that transcribes speech and uses AI to clean up text before inserting it at the cursor.
SelenaCoreSelf-hosted offline smart home hub for Pi/Jetson with local voice (Vosk, Piper, Ollama), modular providers, automations, energy monitoring, multilingual UI — no cloud.
smart-mirrorSelf-hosted open-source voice smart mirror for the Raspberry Pi: control IoT devices by speech, display info services, with plugins, wake-word spotting and remote config.
Sovereign AI WorkstationSelf-hosted AI workstation for macOS running 100% locally with voice, vision, and autonomous agents. Controls apps, writes code, manages Google Workspace.
speech-to-phraseSelf-hosted speech recognition matching only predefined phrases from Home Assistant entities. Runs via Docker or Python, uses Wyoming protocol, supports 20+ languages.
0 comments