A full-stack application that allows users to upload audio files (like call recordings, meetings, or lectures) and search for keywords or phrases within them. The matching audio files will be displayed to user which he can directly listen or read the transcribed text with highlighted matching phrase.
The app automatically transcribes audio into text using OpenAI Whisper and stores it into database for quick, future searches to avoid duplicate processing.
A colleague of mine started to get around 30-35 calls per day from recruiters and with time it became challenging to identify which call number belongs to which company's recruiter after some days. So, created this Audio Query website that allows her to upload many call recording files at once and it get transcribed into the text form. Post that she can use the provided search feature to look for given phrase/keyword, and it'll display the recordings that have those phrases in them, plus user can instantly read the text version of that recoeding instead of listening to it.
Deployed at: Audio Query
i. Home Page
ii. Model Selection
iii. Language Selection
iv. All uploaded audio files and matching searched audio
v. Popup View for details (Transcription is not very accurate as used the base model)
- Backend: Python, Flask, SQLAlchemy, OpenAI-Whisper, Pydub
- Frontend: ReactJS, Vite
- Database: SQLite
- Upload recordings – Supports multiple audio files in one go.
- Automatic transcription – Converts speech to searchable text using Whisper.
- Smart search – Find specific keywords or phrases across all transcribed recordings.
- Audio playback – Play back uploaded recordings directly in the browser.
- Manage recordings – View all recordings and delete unwanted ones.
- Duplicate check – Avoids reprocessing if the same file already exists.
- Preprocessing – Ensures audio is standardized before transcription (mono, 16kHz).
- Multi Language Support – Provided support for multiple languages like English, Hindi, Telugu, French, etc.
- Highlighted Phrase/Keyword – The matching keywords/phrases will be highlighted in the text transcription for easy finding.
- Call centers – Quickly locate calls mentioning “refund,” “complaint,” or “policy.”
- Meetings & lectures – Upload long recordings and search for important discussions later.
- Legal & compliance – Easily review sensitive phrases across large volumes of audio logs.
- Personal productivity – Search through voice notes or interviews without replaying them fully.
Prakash Gupta