---
title: TranscribeAI
emoji: 🎙️
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 5.23.0
app_file: app_gradio.py
pinned: true
license: mit
suggested_hardware: zero-a10g
suggested_storage: small
tags:
  - whisper
  - transcription
  - speaker-diarization
  - faster-whisper
  - indonesian
  - audio
---

# TranscribeAI

100% Free GPU-Accelerated Audio Transcription with Speaker Diarization.

- **Engine**: faster-whisper (CUDA GPU via ZeroGPU H200)
- **Speaker**: MFCC + Agglomerative Clustering
- **Languages**: 99+ (Indonesian, English, etc.)
- **Models**: tiny, base, small, medium, large-v3
- **Output**: SRT, TXT, DOCX