Streamlit AI Speech to Medical Note App
Budget: $10 – $30 USD
Seeking developer to create medical speech-to-note app prototype with Streamlit. The app will transcribe audio of patient encounters into medical visit summaries using Whisper for speech recognition and ChatGPT for language processing.
This is a simple proof of concept that may lead to a more robust product. The goal is to demonstrate the potential for AI to simplify clinical documentation. Success with this project will lead to possible hourly position and enhanced consideration for larger higher budget projects.
Requirements:
- Application must be able to take audio files (and preferably live recording), send it to OpenAI API (Whisper) and return the transcription text, then convert the text using ChatGPT (again with OpenAI API) to return a medical chart note that can easily be copied from the interface for pasting into a medical record.
- Accuracy of Whisper-ChatGPT transcription-to-note conversion should exceed 90% on sample medical dictations
- Ability to integrate with OpenRouter API is a major plus if you are able to do this.
- Intuitive yet simplistic design for minimal training.
Ideal Skills and Experience:
- Experience in developing AI speech-to-text / Streamlit applications.
- Familiarity with medical terminology and healthcare industry.
- Proficiency in using Streamlit for app development.
Please see attached image for a basic visual reference as to the direction we are looking for the sample. Looking for your feedback and suggestions to improve it.
This is a simple proof of concept that may lead to a more robust product. The goal is to demonstrate the potential for AI to simplify clinical documentation. Success with this project will lead to possible hourly position and enhanced consideration for larger higher budget projects.
Requirements:
- Application must be able to take audio files (and preferably live recording), send it to OpenAI API (Whisper) and return the transcription text, then convert the text using ChatGPT (again with OpenAI API) to return a medical chart note that can easily be copied from the interface for pasting into a medical record.
- Accuracy of Whisper-ChatGPT transcription-to-note conversion should exceed 90% on sample medical dictations
- Ability to integrate with OpenRouter API is a major plus if you are able to do this.
- Intuitive yet simplistic design for minimal training.
Ideal Skills and Experience:
- Experience in developing AI speech-to-text / Streamlit applications.
- Familiarity with medical terminology and healthcare industry.
- Proficiency in using Streamlit for app development.
Please see attached image for a basic visual reference as to the direction we are looking for the sample. Looking for your feedback and suggestions to improve it.