🎙️ Speech & Audio AI Repositories

Open-source alternatives for automatic speech recognition (ASR), multi-language transcription, voice cloning, and audio processing.

---

1. OpenAI Whisper

Robust Speech Recognition via Large-Scale Weak Supervision with support for 99+ languages.

• Replaces: Otter.ai ($16/mo) / Descript ($24/mo) / Rev.com ($1.50/min)
• Official Repo: openai/whisper
• License: MIT
• Best For: Accurate speech-to-text transcription and automatic subtitle (.srt / .vtt) generation.

💻 One-Click Install

# Requires ffmpeg installed: (brew install ffmpeg / sudo apt install ffmpeg)
pip install -U openai-whisper

🚀 Instant Transcription Commands

# Transcribe an audio file and output TXT, SRT, VTT
whisper audio.mp3 --model medium

# Transcribe non-English speech and translate directly to English text
whisper foreign_audio.mp3 --model base --task translate

---

2. faster-whisper

Up to 4x faster Whisper reimplementation using CTranslate2 with lower memory usage.

• Replaces: Heavy GPU requirements for speech transcription
• Official Repo: SYSTRAN/faster-whisper
• License: MIT
• Best For: Ultra-fast batch transcription on standard CPUs and small GPUs.

🐍 Copyable Python Snippet

from faster_whisper import WhisperModel

model = WhisperModel("small", device="cpu", compute_type="int8")
segments, info = model.transcribe("meeting_recording.mp3", beam_size=5)

for segment in segments:
    print(f"[{segment.start:.2f}s -> {segment.end:.2f}s] {segment.text}")