React-Flask Speech Transcriber
Budget: $8 – $15 USD
I am racing against a deadline and need a small web app that cleanly transcribes spoken audio to text, end-to-end. The stack is fixed: React for the interface, Flask for the API layer. Here is how I picture the workflow:
• On the React side the user records or drags-and-drops an audio file, sees a progress indicator, then watches the transcript appear in real time.
• Flask receives the audio, feeds it to the speech-to-text engine of your choice (Whisper, Google Speech-to-Text, Vosk—whatever meets the accuracy bar), and streams the text back.
• When the transcript is complete the user can edit inline, copy, download it as a simple .txt or .docx, and—crucially—press “Play” to hear the words read aloud in a natural human-like voice. Please integrate any neural TTS service that delivers that tone; no robotic playback.
Deliverables (all must be met for acceptance)
1. React front end: record/upload component, live transcript view, basic styling.
2. Flask backend: REST endpoints for upload, transcription, natural-voice playback, and health check.
3. Docker-compose file and README so I can spin the stack up with one command.
4. Clear, commented code in a Git repo and a short deployment guide.
Clean architecture, sensible error handling, and quick turnaround are more important to me than fancy UI touches. If you can start immediately, let’s get this shipped.
• On the React side the user records or drags-and-drops an audio file, sees a progress indicator, then watches the transcript appear in real time.
• Flask receives the audio, feeds it to the speech-to-text engine of your choice (Whisper, Google Speech-to-Text, Vosk—whatever meets the accuracy bar), and streams the text back.
• When the transcript is complete the user can edit inline, copy, download it as a simple .txt or .docx, and—crucially—press “Play” to hear the words read aloud in a natural human-like voice. Please integrate any neural TTS service that delivers that tone; no robotic playback.
Deliverables (all must be met for acceptance)
1. React front end: record/upload component, live transcript view, basic styling.
2. Flask backend: REST endpoints for upload, transcription, natural-voice playback, and health check.
3. Docker-compose file and README so I can spin the stack up with one command.
4. Clear, commented code in a Git repo and a short deployment guide.
Clean architecture, sensible error handling, and quick turnaround are more important to me than fancy UI touches. If you can start immediately, let’s get this shipped.