Neura Runtime

Run AI models, execute code, and build apps — all locally on your device.

Detecting...

Local Models

Install Neura Runtime to discover local models.

Start Runtime

neura-runtime start

Stop Runtime

neura-runtime stop

Check status

neura-runtime status

List models

neura-runtime models

Pair device

neura-runtime pair

List Streamlit apps

neura-runtime apps

🧠

Local AI Inference

Run Ollama & LM Studio models (Llama 3, Mistral, Phi, Qwen) on your hardware at $0 cost.

🐳

Code Execution Sandbox

Execute Python in isolated Docker containers with strict sandboxing — network isolation, memory limits, read-only filesystem.

📊

Streamlit App Builder

Agent-generated Streamlit apps run locally on your machine with permanent URLs (no cloud expiry).

🔗

WebSocket Tunnel

Secure tunnel connects your device to Neura Cloud. The agent sends commands, your device executes them.

📁

Local File Management

Files created by agents stay on your device. Full control over data storage and access.

🔒

Device Pairing

Link your device to your Neura account with a 6-digit code. One device per account, revokable anytime.

Install Neura Runtime (desktop app or CLI) on your computer.

Runtime starts a daemon that auto-detects Docker, Ollama models, and LM Studio models on your machine.

A secure WebSocket tunnel connects your device to Neura Cloud. Your browser auto-detects it via localhost:9700.

When you use Neura Agent, it runs AI inference locally, executes code in Docker sandboxes on your device, and hosts Streamlit apps — all at $0 cost.