Whisper
OpenAI's open-source speech recognition model — accurate multilingual transcription, self-hostable.
Last updated by editors:
Our review
Whisper is OpenAI's open-source speech-recognition model: accurate multilingual transcription that you can self-host, with a large ecosystem of wrappers built on top. It has become the open-source cornerstone for transcription and subtitles.
It needs your own compute, does transcription only, and ships without a graphical interface. But if you want free, offline-capable, high-quality transcription that you can wire into your own pipeline, Whisper is close to the default choice.
Pros
Cons
Key features
Best for
Who it's not for
Limits to plan for
- 需自备算力或 API，纯模型无界面
- 只做转录，不做说话人分离等后处理
- 长音频需自行切分与批处理
FAQ
What is Whisper?
OpenAI's open-source speech recognition model — accurate multilingual transcription, self-hostable.
What are the pros and cons of Whisper?
Pros: Open-source & free, Strong multilingual, Runs locally, Rich ecosystem. Cons: Needs compute, Transcription only, No GUI.
How much does Whisper cost?
Open source: Free; API: Pay-as-you-go.
Who is Whisper for?
Best for Speech transcription, Subtitles, Offline recognition, Dev integration.
What are the alternatives to Whisper?
Consider Otter.ai, ElevenLabs, Deepgram.
Pricing
- • Self-hosted
- • OpenAI hosted
Rating
Similar tools
Matched automatically from category and tags, for comparison only.