Whisper
OpenAI’s open-source speech recognition and translation model.
- Stars
- 110.5K
- Forks
- 13.4K
- Open issues
- 171
- Last commit
- 2026-08-31
Live from the GitHub API · updated 2026-10-15
What is Whisper?
Whisper transcribes speech in nearly 100 languages and can translate it into English. The open-source package includes a CLI and Python API with model sizes from tiny to large, plus the faster “turbo” model.
Key features
- Multilingual transcription and translation to English
- Model sizes from tiny to large-v3 and turbo
- SRT/VTT subtitle output
- Language detection
System requirements
- Platforms
- Windows · macOS · Linux
- GPU
- Optional; NVIDIA recommended for large models
- VRAM
- ~1 GB (tiny) to ~10 GB (large)
- Python
- 3.8–3.11
- Notes
- Requires ffmpeg on PATH.
How to install Whisper
Summarized from the project’s README. Projects change quickly — check the GitHub page for the latest instructions.
pip
- 1Install ffmpeg (e.g. brew install ffmpeg, sudo apt install ffmpeg or choco install ffmpeg), then Whisper.
pip install -U openai-whisper - 2Transcribe a file.
whisper audio.mp3 --model turbo --output_format srt
Strengths
- Excellent accuracy across languages
- MIT license
Limitations
- Slow on CPU
- Can hallucinate on silence
Whisper alternatives compared
| Software | Stars | License | Platforms | VRAM |
|---|---|---|---|---|
| Whisper OpenAI’s open-source speech recognition and translation model. | 110.5K | MIT | ~1 GB (tiny) to ~10 GB (large) | |
| faster-whisper Whisper reimplemented on CTranslate2 — up to 4× faster with less memory. | 25.9K | MIT | — | |
| whisper.cpp Whisper in plain C/C++ — fast on CPUs, Apple Silicon and mobile. | 54.4K | MIT | — | |
| WhisperX Fast transcription with word-level timestamps and speaker diarization. | 24.6K | BSD-2-Clause | — | |
| Coqui TTS Deep-learning text-to-speech toolkit, including the XTTS voice cloning model. | 46.1K | MPL-2.0 | — | |
| GPT-SoVITS Few-shot voice cloning and TTS with a full training WebUI. | 62.8K | MIT | 6 GB+ |
FAQ
Is Whisper free?
Yes — Whisper is open-source software released under the MIT license. Check the license (and any model weight licenses) before commercial use.
Does Whisper run on Windows, Mac and Linux?
Whisper supports: Windows, macOS, Linux.
What hardware do I need for Whisper?
Optional; NVIDIA recommended for large models. VRAM: ~1 GB (tiny) to ~10 GB (large)
How do I install Whisper?
1. Install ffmpeg (e.g. brew install ffmpeg, sudo apt install ffmpeg or choco install ffmpeg), then Whisper. 2. Transcribe a file.
What are the best Whisper alternatives?
Popular alternatives include faster-whisper, whisper.cpp, WhisperX, Coqui TTS. See the comparison table on this page.