Wraith Signal Copier Vision

Everything in Signal Copier Base, plus image and chart-screenshot signal parsing. Channels that post setups as screenshots instead of text get parsed too, through a pluggable, provider-substitutable vision engine. Defaults to OpenAI, but you can point it at any OpenAI-compatible vision provider (NVIDIA build.nvidia.com, OpenRouter, and others) with your own API URL, key, and model or run native Gemini or Anthropic engines instead. Bring your own vision model and key: your signals never leave your box, and Vision parsing costs us nothing to run, so it costs you nothing beyond your own API usage. Self-hosted, one-time license, no relay fee.