This is a submission for the AssemblyAI Challenge : No More Monkey Business.
What I Built
The project Athene AI is a digital journal note powered by AssemblyAI's speech-to-text model. It allows user to record journal entries and it auto transcribes it, generate summaries, proper speech segmentation, does mood analysis and auto create tags for characterizing the note.
Demo
The app is live and can be used here:
Athene AI
The code is also available here:
Frontend
Backend
Video Demo
Journey
I figured that I would journal more if i can just talk to something and it gets transcribed to readable words, organized by tags and also have a bit of categorization that will help me remember what the mood or environment was when I entered that journal.
With LeMUR, when i transcribe the audio of the journal, I used the inbuilt summary method to get summary. I also created two task that is processed by LeMUR: to generate a title for the journal entry and to auto tag it based on categories it fall into.
Note: This submission also qualifies for the Sophisticated Speech-To-Text prompt. Using Universal-2, AssemblyAI’s latest and most accurate speech-to-text model, I built the transcription engine for the journal app. It processes the journal entries smoothly while maintaining a good user experience.
The whole engine runs on Cloudflare workers deployment with Hono web framework. It also uses KV for storage of the transcribed data.
Top comments (0)