READABLE

ReadAble is an accessibility utility that uses OCR, text-to-speech, and translation to make text easier to access. It is built for users who struggle with visual text, dense pages, unfamiliar language, or reading-heavy workflows.

A lot of information is technically available but practically hard to use. Text inside images cannot be copied, long passages can be tiring to read, and language differences add friction. Accessibility is not a bonus feature; it is the difference between usable and decorative.

The tool uses a JavaScript-based frontend flow with API integrations for OCR, text-to-speech, and translation. The user provides visual or written input, the OCR layer extracts text where needed, and the result can be read aloud or translated based on the selected workflow.

The state model is kept simple: input, extracted text, transformed output, and user action. This keeps the UI understandable and avoids turning a utility into a settings museum.

ReadAble is built around practical accessibility operations.

The difficult part is keeping the extracted text clean enough for the next step. OCR mistakes affect speech and translation quality, so each layer depends on the previous one behaving reasonably well. Pipeline discipline matters even in a small tool.

I kept the product focused on a small number of transformations instead of adding unrelated features. OCR, speech, and translation were chosen because they solve the same accessibility problem from different angles: seeing, hearing, and understanding.

The prototype provides a clear accessibility workflow and a base for future improvements like better OCR cleanup, saved outputs, language presets, and mobile-first capture support.