OpenClaw integration

Your phone talks to any AI model. One command.

The AutoYou OpenClaw plugin bridges your phone directly into OpenClaw's full AI pipeline. Every model, every tool, every agent — accessible from chat on your phone. Auto-start included.

Auto-start Browser proxy Security modes Cloud pairing TOTP 2FA All AI models

What the plugin does

AutoYou + OpenClaw — how they connect

OpenClaw is the AI gateway running on your computer. The AutoYou plugin sits inside it and bridges the mobile app to the full OpenClaw AI pipeline via a live WebRTC connection.

Message routing

Chat routes to any AI model

Messages from the AutoYou app arrive via Server-Sent Events (/api/v1/events). The plugin forwards them to OpenClaw's AI pipeline — which can be Claude, GPT-4, Gemini, a local Ollama model, or any custom agent.

Auto-start

Plugin manages autoyou_lib for you

Set "autoStartLib": true in the plugin config and OpenClaw spawns autoyou_lib automatically on startup — with all your security, proxy, and cloud settings passed through. No manual server management needed.

AI model provider

AutoYou registers as a model provider

The plugin registers an autoyou/default model provider in OpenClaw. When a message comes from the AutoYou app, OpenClaw routes it through the AI pipeline and streams the response back to the app via the DataChannel.

Installation

Install in one command

1

Install the plugin

openclaw plugins install @openclaw/autoyou

Verify it's active:

openclaw plugins list
# autoyou   0.2.0   active
2

Install Python dependencies

# Core (no voice)
pip install fastapi uvicorn aiohttp pydantic cryptography aiortc av

# Voice support
pip install openai-whisper sounddevice numpy scipy

# TOTP (Secure Professional mode)
pip install pyotp
3

Add the plugin config to openclaw.json

{
  "plugins": {
    "autoyou": {
      "autoStartLib": true,
      "port": 8099,
      "authPort": 8098,
      "securityMode": "open",
      "proxyEnabled": true,
      "proxyTarget": "http://127.0.0.1:8856"
    }
  }
}
4

Start OpenClaw

openclaw gateway run

The plugin starts autoyou_lib automatically (if autoStartLib is true), registers the autoyou/default model provider, and begins listening for SSE events from the server.

Configuration reference

All plugin options

Set these fields inside "plugins": { "autoyou": { ... } } in your openclaw.json.

Key Type Default Description
autoStartLib boolean false Spawn and supervise autoyou_lib automatically on plugin start
libUrl string http://127.0.0.1:8099 Base URL for the autoyou_lib REST API (used when autoStartLib is false)
port number 8099 Port for the autoyou_lib internal REST API
authPort number 8098 Port for the public /auth and /signal endpoints
securityMode enum "open" Security mode: "open", "secure", or "secure-pro"
totpSecret string Base32 TOTP secret for secure-pro mode
proxyEnabled boolean false Enable the DataChannel HTTP proxy
proxyTarget string http://127.0.0.1:8856 Local service URL to forward browser proxy requests to
cloudApiUrl string AutoYou Cloud API URL for cloud relay pairing (paid tier)
cloudDeviceToken string Device token for AutoYou Cloud (paid tier)
agentUrl string External AI agent URL override (bypasses OpenClaw pipeline)
pythonBin string "python3" Python binary path for auto-starting autoyou_lib
logLevel string "info" Log level for autoyou_lib: debug, info, warning, error

Security

Security modes via the plugin

The plugin passes securityMode (and totpSecret for Secure-Pro) to autoyou_lib on startup. The app adapts automatically.

Open

Default — WebRTC security

{
  "securityMode": "open"
}

Standard WebRTC DTLS/SRTP. No additional configuration needed.

Secure Professional

TOTP 2FA on every reconnect

{
  "securityMode": "secure-pro",
  "totpSecret": "YOUR_BASE32_SECRET"
}

TOTP secret shared with the app at first pairing. Every reconnect requires a valid TOTP code.

Commands

Plugin slash commands

These commands are available in the OpenClaw interface after the plugin is installed.

/autoyou-status

Shows the current status of the AutoYou connection: server running/stopped, security mode, TOTP enabled, browser proxy state, and AutoYou Cloud enrollment. Example output:

AutoYou status
  Server:        running (port 8099)
  Security:      secure-pro (TOTP enabled)
  Browser proxy: enabled → http://127.0.0.1:8856
  Cloud pairing: enrolled (server_id: abc123)
/autoyou-enroll [cloud_url] [device_token]

Enroll the running server with AutoYou Cloud at runtime — no restart required. If cloudApiUrl and cloudDeviceToken are already in the plugin config, call /autoyou-enroll with no arguments to use them.

# Use config values
/autoyou-enroll

# Or pass explicitly
/autoyou-enroll https://cloud.autoyou.me YOUR_DEVICE_TOKEN

Paid tier

AutoYou Cloud pairing

Paid subscribers can connect via AutoYou Cloud — eliminating tunnelmole and making AutoPair work reliably from any network. The cloud relays only the WebRTC signaling; your data flows P2P once the connection is established.

How it works

Signaling relay, P2P data

The server registers with AutoYou Cloud on startup and long-polls an inbox for incoming AutoPair offers from the app. When an offer arrives, the server generates a WebRTC answer and posts it back via the cloud. The cloud relays it to the waiting app.

Once the WebRTC handshake completes, the cloud is no longer involved. Chat, browser requests, and voice all flow directly P2P — the cloud never sees your messages.

App → Cloud → Server: autopair_offer
Server → Cloud → App: autopair_answer
App <──────────────────────────> Server
       (direct P2P — no cloud)

Three ways to enroll

Choose your setup method

1. Plugin config (auto-enroll on start)

{
  "cloudApiUrl": "https://cloud.autoyou.me",
  "cloudDeviceToken": "YOUR_DEVICE_TOKEN"
}

2. Runtime command (no restart needed)

/autoyou-enroll https://cloud.autoyou.me YOUR_TOKEN

3. CLI flag (standalone server)

python -m autoyou_lib \
  --cloud-api-url https://cloud.autoyou.me \
  --cloud-device-token YOUR_TOKEN

Check enrollment: Run /autoyou-status — it shows Cloud pairing: enrolled (server_id: ...) when active.

Full picture

AutoYou + OpenClaw architecture

┌─────────────────────────────────────────────────────────┐
│               AutoYou App (iOS / Android)               │
│  Chat · Browser · Voice · Notes · Agent Frontends       │
└──────────────────────┬──────────────────────────────────┘
                       │
         ① OTP/AutoPair │ (tunnelmole or AutoYou Cloud)
         ② WebRTC P2P   │ (SRTP audio + SCTP DataChannel)
                       │
┌──────────────────────▼──────────────────────────────────┐
│              autoyou_lib   (Python, your machine)        │
│  Port 8098: /auth /signal  (public, via tunnelmole)      │
│  Port 8099: REST + SSE     (internal)                    │
│                                                          │
│  DataChannel → HTTP proxy → proxy_target (port 8856)    │
│  SRTP track  → Whisper STT → AI → TTS → app             │
└──────────────────────┬──────────────────────────────────┘
                       │ SSE /api/v1/events
                       │
┌──────────────────────▼──────────────────────────────────┐
│          OpenClaw autoyou plugin  (Node.js)              │
│  Routes chat → OpenClaw AI pipeline (port 8856)          │
│  Fallback   → autoyou_lib /v1/chat/completions           │
│  Registers  → autoyou/default model provider             │
└──────────────────────┬──────────────────────────────────┘
                       │
┌──────────────────────▼──────────────────────────────────┐
│              OpenClaw gateway  (port 8856)               │
│  Claude · GPT-4 · Gemini · Ollama · Any custom agent    │
└─────────────────────────────────────────────────────────┘

Get connected

Install once. Chat from anywhere.

One plugin command, one Python server, and your phone has direct access to any AI model running on your computer. No account needed. Free.