nexting-hermes-bridge

Connect Nexting hardware to any OpenAI-compatible local AI (Hermes Agent, Ollama, vLLM, LM Studio, etc.)

Prerequisites

• Node.js v18+
• A local AI with OpenAI-compatible API running (e.g. Hermes Agent)
• Nexting iOS app installed and logged in

Quick Start

1. Install

npm install -g nexting-hermes-bridge

Or run directly without installing:

npx nexting-hermes-bridge

2. Start your local AI

Hermes Agent:

Make sure your ~/.hermes/config.yaml has the API server enabled:

platforms:
  api_server:
    enabled: true
    host: "127.0.0.1"
    port: 8642

Then start the gateway:

hermes gateway run

Verify it's running:

curl http://localhost:8642/v1/models

Ollama (alternative):

ollama serve
# Default endpoint: http://localhost:11434

3. Pair with your iPhone

nexting-hermes-bridge login

This will:

• Create a relay instance on the Nexting cloud
• Display a 6-digit pairing code and QR code
• Wait for you to enter the code in the Nexting iOS app

Open the Nexting iOS app → Settings → select "MyHermes" mode → enter the pairing code.

4. Start the bridge

nexting-hermes-bridge start

The bridge will:

• Connect to the Nexting cloud relay
• Forward messages from your iPhone to your local AI
• Stream responses back to your iPhone
• Handle device tool calls (calendar, reminders, etc.)

Done. Speak into your Nexting device and your local AI will respond.

Options

# Use a different AI endpoint (default: http://localhost:8642)
nexting-hermes-bridge start --endpoint http://localhost:11434

# Specify model name
nexting-hermes-bridge start --model llama3

# Specify API key (if your endpoint requires one)
nexting-hermes-bridge start --api-key sk-xxx

# Check current config
nexting-hermes-bridge status

# Clear config and start over
nexting-hermes-bridge logout

How it works

Nexting Device → BLE → iPhone (STT) → Cloud Relay → nexting-hermes-bridge → Local AI API
                                                   ↑
                                            Your computer

The bridge runs on your computer. It connects to the Nexting cloud as a relay endpoint (same infrastructure as MyOpenClaw). The cloud only tunnels encrypted messages — your AI conversations stay between your iPhone and your local machine.

Developing

cd bridge
npm install

# Run in dev mode
npx tsx src/index.ts login
npx tsx src/index.ts start

# Type check
npx tsc --noEmit

# Build
npm run build

Config

Config is stored at ~/.nexting/config.json:

{
  "relayToken": "...",
  "relayUrl": "wss://api.nexting.ai",
  "endpoint": "http://localhost:8642",
  "model": "hermes-agent",
  "subdomain": "h-xxxxxxxx"
}

Device state (skills, context) is persisted at ~/.nexting/device-state.json.