← Back to work
Speakscribe
Transcription tools hand you a wall of text. But never let you actually ask it anything.
PythonFastAPIOpenAI APIReact
The problem
Audio transcription tools are often locked behind clunky UIs, and rarely let you interrogate what was said.
Diagnosis
The interesting part wasn’t transcription alone. It was pairing accurate speech-to-text with a conversational layer so users could query and interact with the content.
What I built
- FastAPI backend wrapping OpenAI transcription and chat.
- React front-end for upload, playback, and conversation.
- Natural-language Q&A over the generated transcript.