Sanket: The App That Turns Speech Into Indian Sign Language

A Flutter-based mediator for DHH communication that pairs 3D sign animations with a simple open-source workflow for translating speech into ISL.

8 min read View on GitHub More from David27322

A speaker faces a Deaf or Hard of Hearing user across a clean, staged scene. Between them, a phone projects a sequence of 3D sign language gestures that act like a bridge, not a subtitle strip. The image explains Sanket as a mediator between spoken language and Indian Sign Language.
Sanket is interesting because it does not just display words. It tries to mediate communication through animated sign language output.
Key Takeaways

Sanket is not a chat app with a sign-language skin. Its job is narrower and more interesting: it tries to translate speech into Indian Sign Language using animated 3D output, so the result feels like mediation instead of annotation.

That matters because the gap is not just linguistic. It is also visual, cultural, and technical. A tool like this has to make signs legible enough to be useful, while keeping the interaction simple enough to run on a phone.

The app’s main job is to make speech visible in ISL

The repository centers on a mobile experience built with Flutter. That is a sensible choice for a communication tool, because the app needs to run cleanly on phones and keep the UI light.

The distinctive part is the output layer. Instead of relying only on text or static graphics, Sanket uses 3D animations created in Blender to represent Indian Sign Language.

Software developer passionate about building efficient, scalable solutions. Always learning, always coding.

David Sharma, Author/Maintainer · David27322 (David Sharma) - GitHub
A compact pipeline shows speech entering one side of the system and emerging as a sequence of signed gestures on the other. The left side fragments language into smaller units, while the right side reassembles them into a coherent animation stream. The illustration explains how Sanket turns input into signed output.
The important transition is from spoken language to a signed visual sequence. Sanket’s value sits in that conversion layer.

Why the visual layer matters more than a text transcript

For hearing users, text often feels like the end state. For a sign-language mediator, text is only an intermediate form. Sanket’s 3D avatar approach makes that distinction visible.

That is also why the project feels closer to communication infrastructure than to a demo. It is not trying to maximize novelty. It is trying to preserve meaning across two different language modes.

This diagram shows Sanket as a translation pipeline, with the hard boundary sitting between spoken input and signed visual output.

What the stack tells you about the project

The stack is intentionally pragmatic. Flutter gives it a cross-platform mobile shell, while Blender handles the asset creation side of the system. That division keeps the application focused on delivery instead of overengineering the animation layer.

The repository also suggests a student-project origin. That is not a weakness here. In fact, it explains the shape of the work: a focused prototype that solves one communication problem clearly, rather than a sprawling platform trying to do everything.

DimensionSanketTypical sign-language app
Primary goalMediating speech into ISLShowing text or basic translation output
Visual output3D sign animationsStatic icons or captions
Mobile stackFlutterOften web-first or mixed
Asset creationBlender-produced animation assetsGeneric UI assets
User experienceCommunication bridgeInformation display
Project feelFocused prototypeFeature bundle

The most useful comparison is not against a generic consumer app. It is against any tool that stops at transcription. Sanket is trying to move one step further and render the message in a form that can actually be received as sign language.

The open-source value is in the narrowness

A lot of open-source apps try to broaden their scope early. Sanket does the opposite. It chooses a specific language context, a specific interaction model, and a specific output format.

That narrowness is what makes it interesting. A communication tool earns trust by being precise about who it serves and what it translates.


Sources