Whisper OpenAI Online: Turn Any Recording Into Text
Upload a recording and get it back as written text with Whisper OpenAI online. The model understands speech in 98 languages, and the finished transcript stays right in your chat. You can then switch to another AI model and keep working on the text without starting over.
Whisper transcription online starts at no cost on Cabina.AI. The free plan comes with tokens for several transcriptions, and you don't pay for each model separately. One account gives you Whisper plus GPT, Claude and the other top models, all in one place and one chat.
What You Can Do With Whisper
One model covers the speech jobs most people bring to it. Whisper speech recognition is built for general use, and the same model can transcribe a recording, translate it and tell which language it hears.
Transcribe Speech in Many Languages
Turn a lecture, a call or an interview into readable text. It works for English and for the other supported languages, so you don't need a separate tool for each one.
Translate Speech Into English
Translation is a built-in skill of the Whisper model. It can take speech in another language and write it out directly as English text, with no separate translation pass needed.
Detect the Spoken Language
You don't need to name the language before you upload. Whisper identifies the spoken language by itself, which is handy when your recordings come from different countries.
Trained on 680,000 Hours of Audio
Whisper AI transcription comes from a model trained on 680,000 hours of internet audio with matching transcripts. Even so, it can add words nobody said, so double-check names and figures.
Go Further With Other AI Models
Whisper translates only into English. For any other target language, ask a text model in the same chat to translate the transcript. It can also summarize or rewrite the text for you.
Which Whisper Version to Pick
You'll find three versions under the Whisper name. Each OpenAI Whisper model turns speech into text, but they differ in extra abilities and in how many tokens one message costs, so pick by the job in front of you.
| Version | What It Can Do | Tokens per Message |
|---|---|---|
| Whisper v2-Large | Transcribes speech and can translate it into English. It's also the only one of the three models with timestamp support. | 11.92 |
| GPT-4o Transcribe | Built on GPT-4o and made for transcription only, with no translation into English. | 17.51 |
| GPT-4o Mini Transcribe | Built on GPT-4o Mini for transcription only, with the lowest token cost of the three. | 11.49 |
How to Use Whisper on Cabina.AI
If you're wondering how to use OpenAI Whisper here, it takes four steps. The last one goes beyond a plain transcription app, because your text stays in a chat where other AI models can work with it.
Create a Free Account
Sign up for the free plan to get free Whisper transcription online. Its tokens work for Whisper and for every other model too.
Pick Whisper and a Version
Choose Whisper from the audio models, then select Whisper v2-Large, GPT-4o Transcribe or GPT-4o Mini Transcribe for the task.
Upload Your File
Add the audio or video file you want as text. Recording from a microphone isn't available, so save your recording as a file first.
Read and Reuse the Transcript
Read the text Whisper returns. Then switch to GPT or Claude in the same chat and ask for a summary or a cleaner rewrite.
Transcribe First, Then Rewrite With Another AI Model
Use Whisper AI for transcription, then let a text model work on the result with no copy and paste. When you switch models in one conversation, the chat history stays, so ChatGPT, Claude, Gemini or another text model already sees the transcript. Ask it for a meeting summary, a list of action items or a blog article built from an interview.
Want to see which model writes the better summary? Compare Mode sends one request to two models and shows the answers side by side, so you keep the stronger one. You can also have the finished text read aloud with OpenAI text-to-speech. With an OpenAI Whisper transcription tool, a text model and a voice model in one place, you go from a recording to a polished script and back to audio.
Your recording
Audio file
Whisper AI
Transcription
ChatGPT, Claude, Gemini
Chat history stays
ChatGPT
Claude
OpenAI text-to-speech
Read aloud
Alternatives to Whisper
OpenAIAudio 
ElevenLabs 
CosyVoice 
ElevenLabsScribe 
GeminiTts 
SeedAudio 
MiniMaxAudio Use Whisper in Claude, ChatGPT or Cursor Through MCP
Developers can call Whisper for transcription right from Claude, ChatGPT, Cursor or Windsurf. Cabina.AI offers API access and an MCP connection. MCP stands for Model Context Protocol, and it lets these apps use over 55 AI models, Whisper among them, through one connection. Setup takes a single click or one terminal command with your personal API key, and all models are billed together. See how to connect Cabina.AI over MCP.
Claude
ChatGPT
Cursor
Windsurf
Cabina.AI
55+ AI models
Whisper Pricing and Free Plan
OpenAI Whisper pricing here works in tokens. Each message costs a set number of tokens for the version you choose, and every plan shows how many messages its tokens cover.
| Plan | Price | Tokens | Whisper v2-Large | GPT-4o Transcribe | GPT-4o Mini Transcribe |
|---|---|---|---|---|---|
| Free | $0.00 | 50 | 8 | 5 | 8 |
| AI | $4.99 per month | 450 | 37 | 25 | 39 |
| Agent | $19.99 per month | 1,000 | 83 | 57 | 87 |
| Agent Pro | $49.99 per month | 2,100 | 176 | 119 | 182 |
| Pay as You Go | $20 | 1,800 | 151 | 102 | 156 |
Free$0.00
AI$4.99 per month
Agent$19.99 per month
Agent Pro$49.99 per month
Pay as You Go$20
Tokens you don't spend on transcripts also work for text, image and video models. The Pay as You Go row is a one-time top-up for anyone who doesn't want a subscription.
Whisper FAQ
What Audio Files Can I Use With Whisper?
You can upload audio in MP3, MPEG, M4A, AAC, AIFF, FLAC, OGG, WAV and WEBM. Video files work too: MP4, MPG, MOV, AVI, WMV, FLV and 3GP are accepted, so you can transcribe a phone video or a screen recording as it is.
Is Whisper Free to Use?
Yes, in two ways. OpenAI released the Whisper code and model weights under the MIT License, so the software is free to download and run yourself. If you'd rather use OpenAI Whisper free online, the free plan on Cabina.AI includes tokens for your first transcriptions.
How Accurate Is Whisper?
It depends mostly on the language. OpenAI publishes error rates for each language, and accuracy follows the amount of training audio a language had, so less common languages come out weaker. The model may also repeat a phrase in its output, so give important transcripts a quick check against the recording.
Can Whisper Label Different Speakers?
No. Whisper voice to text returns the spoken words without marking who said each line. OpenAI offers speaker labels through a separate model, gpt-4o-transcribe-diarize, and that model isn't one of the three versions you can pick here.
Can I Run Whisper on My Own Computer?
Yes. The open-source release comes in six sizes, from tiny, which needs about 1 GB of video memory, to large, which needs about 10 GB. Tiny runs about 10x faster than large, and turbo comes close at about 8x. A weaker graphics card limits you to the smaller sizes.
How Much Does Whisper Cost per Minute?
On the OpenAI API, the estimated Whisper cost per minute is $0.006, the same as GPT-4o Transcribe, while GPT-4o Mini Transcribe costs $0.003. Those figures are OpenAI rates for developers. Cabina.AI plans charge in tokens instead, so compare the version you pick with the tokens in your plan.
Does Whisper Work in Real Time?
Not out of the box. The open Whisper models can't transcribe live speech by default, so Whisper web transcription on Cabina.AI starts from a finished recording that you upload. GPT-4o Transcribe and GPT-4o Mini Transcribe do support live audio through the OpenAI Realtime API for developers.
How Do I Get OpenAI Whisper?
Developers can get the Whisper speech-to-text model from the official OpenAI GitHub page and install it on their own machine. To skip the setup, sign up for Cabina.AI and pick Whisper from the audio models right in your browser.
What Is Better Than Whisper for Transcription?
That depends on your audio, so a quick test gives the clearest answer. Your account also includes Cabina Transcriber and ElevenLabs Scribe transcription among the audio models. Run the same file through each, then decide whether a Whisper transcription service or another model suits your recordings.
Can I Use Whisper Without a Subscription?
Turn Your Next Recording Into Text
Sign up free, upload your first file to Whisper, then keep shaping the text with GPT, Claude and every other model in the same Cabina.AI account.
