-
-
User can upload a voice memo. The audio file will be transcribed.
-
A summary of the transcript is used to pre-fill the thought record exercise. The user are encouraged to further expand on it.
-
User is guided to fill out evidence for and against their thoughts, and develop a coping strategy.
-
At the bottom, user can export the exercise for their personal use.
Inspiration
Pandemic, war, economic stress, career, and personal life changes can greatly affect our mental health. In observance of mental health awareness month, I want to make something that helps people maintain balanced thoughts.
I've always find thought records used by cognitive behavioural therapist helpful in restructuring intrusive thoughts and developing coping strategy. So I made Thought Diary to help people log their thoughts and track their progress.
What it does
Thought Diary transcribes user's voice memo describing a stressful scenario. From the transcript, it separates the situation, emotions, and thoughts out of the user's narrative. Later, the user is prompted to examine their thoughts, and form coping strategies. When the exercise is completed, the user can generate a report of their progress.
It can also be used as an educational resource for people learning about cognitive behavioural therapy.
How we built it
- Used AssemblyAI's api to transcribe audio to text.
- Used OpenAI's api to summarize the situation, emotions, and thoughts in the transcript.
- Used streamlit and python to build the web app.
Challenges we ran into
- The transcript is not perfectly accurate.
- The summarization of the transcript might have weird layout that cannot be parsed correctly by existing code.
- Not sure how to create multiple web pages with streamlit
- The information being handled is sensitive. Need to find out how to protect people's privacy.
Accomplishments that we're proud of
Thought Diary is a fully functioning app. I've demoed it to friends, and have them tried out the thought record exercise. It feels awesome to promote mental health awareness among loved ones!
What we learned
- AssemblyAI's api
- streamlit cloud
What's next for Thought Diary
- Add an alert when user uploads an invalid audio file
- I'd like to see more people trying out Thought Diary and get their feedback.
- I'd like to learn what do mental health professionals think of the tool
Built With
- assemblyai
- openai
- python
- streamlit
Log in or sign up for Devpost to join the conversation.