Code Up

Inspiration

Programming is one of the most powerful skills today, yet it remains largely inaccessible to people with visual impairments. Most modern IDEs are designed around visual interfaces such as cursors, syntax highlighting, and graphical debugging tools, which makes them difficult to navigate using screen readers. We wanted to rethink what a development environment would look like if it were designed specifically for blind programmers from the beginning. Code_Up was inspired by the idea that programming should depend on logic and creativity rather than eyesight, and that accessibility should be treated as a fundamental design principle rather than an optional feature.

What it does

Code_Up is a voice-controlled Python IDE designed for visually impaired developers. It enables users to write, edit, execute, and debug Python code through voice commands and audio feedback.

The system integrates speech recognition for code input and text-to-speech feedback to read code, errors, and navigation information aloud. Code_Up also includes an AI coding assistant powered by Gemini, which helps generate code snippets, explain errors, and assist with debugging. Additional developer tools such as syntax highlighting, linting, automated test case generation, and command palette navigation help provide functionality comparable to modern development environments.

By combining voice interaction, AI assistance, and accessibility-first design, Code_Up allows visually impaired users to independently develop and run Python programs.

How we built it

Code_Up was primarily built using Python with a PyQt5-based interface designed to support accessibility and voice interaction.

Speech recognition technology is used to capture spoken commands and convert them into code input. Text-to-speech systems provide auditory feedback so users can hear their code, navigation changes, and error messages.

For AI assistance, the system integrates the Google Gemini API through the Google GenAI Python SDK. Gemini is used to generate code suggestions, explain programming concepts, assist with debugging, and support natural language programming queries.

The architecture separates the voice interface, AI assistant, code editor, and execution engine into modular components, allowing the platform to remain extensible for future improvements.

Challenges we ran into

One of the biggest challenges was redesigning typical IDE interactions so they work without relying on visual cues. Concepts such as cursor position, indentation structure, and syntax highlighting had to be translated into meaningful audio feedback.

Speech recognition also presents reliability challenges because programming syntax is extremely precise. Small transcription errors can lead to invalid code, so we had to design voice commands and editing workflows that allow users to correct mistakes efficiently.

Another challenge was integrating AI assistance in a way that provides useful support without overwhelming the user with excessive spoken information.

Accomplishments that we're proud of

We are proud that Code_Up demonstrates how a programming environment can be designed around accessibility rather than retrofitted to support it. The platform combines voice interaction, AI assistance, linting tools, and automated testing into a single development environment tailored for visually impaired users.

Successfully integrating voice coding with AI-powered development assistance represents a meaningful step toward making programming more inclusive.

What we learned

Building Code_Up taught us how important human-centered design is when building developer tools. Accessibility challenges forced us to rethink assumptions about how programmers interact with software.

We also learned how to integrate speech recognition, text-to-speech systems, and large language models into a cohesive workflow. Balancing usability, reliability, and responsiveness was essential to making the system practical.

What's next for Code_Up

Future development will focus on improving usability and expanding capabilities. Planned features include natural-language-to-code generation, smarter debugging assistance, improved voice navigation, and an interactive tutorial system to help visually impaired beginners learn programming.

We also plan to expand support beyond Python and introduce collaborative features so visually impaired developers can more easily work with teams.

Our long-term goal is to develop Code_Up into a fully accessible development platform that empowers visually impaired programmers to build software independently.

Built With

• automatedtestgen
• css
• googlegeminiapi
• html
• javascript
• lintingtools
• pyqt5
• python
• speechrecognition
• text-to-speech
• voicecommand