SignType V6 — signing workspace

Sign to text with local camera tracking: trainable word signs, fingerspelling, your own models, and a two-palm SEND gesture.

SignType V6 turns your camera into a personal workspace for signing, typing, learning and connecting. Click a supported text field, start the camera and sign using your trained vocabulary or selected recognition mode. Recognized text appears automatically in the field, with a copy retained in your draft. Features • On-device hand, face and body tracking with adjustable camera preview and framing guidance • Trainable word signing: record examples of your own gestures and build a personal vocabulary, with countdowns, visible progress and optional spoken feedback • Experimental ASL fingerspelling, automatic page typing and draft review • Optional speech output to read your recognized text aloud • Sign-triggered visual effects: map a trained gesture to a heart, star, ring or sparkles—with text or as an effect alone • Private one-to-one video calls with text chat, optional recognition captions, shared practice prompts and sign-design messages • Separate controls for sharing your camera, microphone, captions and effects • Bring your own recognition model: load supported ONNX models or trained word models, or configure a compatible translator endpoint • Connect your preferred AI model through supported OpenAI, Anthropic, Gemini or custom Chat API formats, using your own API access • Ask AI about text or a reviewed still image, then copy, speak or add its response to your draft • Optional SEND gesture, plus backspace, word deletion, newline and undo controls • Guided walkthrough and dedicated Write, Train, Call, Connect and Help tabs Privacy and control Local camera tracking and local recognition run on your device. Training examples are stored locally. Calls transmit the media and messages you choose to share through an encrypted WebRTC connection. Optional provider features send only the inputs you approve to your configured service. SignType has no call-recording feature. Compatibility Confirmed working on macOS and Windows, with and without dedicated GPUs. Linux, including Omarchy, is a target platform; physical-device testing remains pending. Supports many ordinary text inputs, textareas and compatible rich-text editors. Protected browser pages and some specialized editors cannot receive text. SignType is an evolving accessibility project. Recognition accuracy varies, and experimental fingerspelling and personal vocabulary recognition are not fluent conversational ASL translation. Continuous translation requires a compatible recognition model or service. General AI connections do not automatically provide ASL translation or SL2T access. External services may require separate accounts and fees.