System

Settings

GPU provider, VRAM management, and cloud API configuration.

Colab GPU pool

Clips generate in parallel across all alive Colab GPUs.

Offline

0 active Colab GPUs. Start a profile on the Playwright page and wait for it to reach ALIVE.

Manage

Voice AI Service

http://127.0.0.1:5114 · local

Offline

Idle Timeout

Model auto-offloads after 600s idle

600s

GPU & VRAM

No GPU detected

Models on disk

D:\AI\_VoiceMaster

No models detected

Cloudflare Tunnels

Auto-provision tunnels on scorpion.codes when you add Colab GPU profiles

Not verified

API TokenAccount Cloudflare Tunnel Edit + Zone DNS Edit

DNS (Global API Key)

API tokens often cannot use the DNS API even with all policies — a Cloudflare limitation. Add account email + Global API Key here (My Profile → API Keys). Tunnels use the API token; DNS records use the Global Key.

Account emailGlobal API Key

Account IDZone ID (optional)Auto-resolved from domain if emptyTunnel domainSubdomain prefixe.g. tts → tts-gpu-1.scorpion.codes

Paste API token + Global API Key, then Verify. DNS uses Global Key; tunnels use API token.

ElevenLabs (Cloud)

API keys for cloud TTS fallback

API KeysOne per line. Keys are rotated automatically.

Voice IDModel