Voxtral Studio

Voxtral Studio

Turns speech into text, text into speech and a voice sample into a reusable cloned voice, using Mistral AI's Voxtral models through your own Mistral API key. For people who need transcripts or voice-overs.

Made by codexxx on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: Voxtral Studio is a Gradio app built on Mistral AI's Voxtral models. It does three jobs. It turns speech into text, text into speech, and a short voice sample into a reusable voice you can call by an ID. It needs a paid Mistral API key, which you add as a secret in the Space settings.

Who it suits: It suits people who need transcripts or spoken audio and already have a Mistral account, and developers who want a small app that wraps the Voxtral models.

Industry | Who uses it | How they use it | Example
Media | Podcasters and video editors | Turn a recording into text, or a script into audio | Transcribe an interview in one of the supported languages
Marketing | Content teams | Make a voice-over from written copy | Save the audio as mp3 for a short clip
Accessibility | Publishers | Create spoken versions of written pages | Read an article aloud in a clear voice
Software | Developers | Read how the app calls the speech models | Copy the Space and add another feature

You can skip it if you have no Mistral API key or no billing set up, since the text-to-speech part has no free tier.

Where to use it:

• A web browser: open the Space page on Hugging Face, with your key set as a secret.
• Windows, macOS or Linux: run a copy with Python 3.11.

Use case | Who | What they get back
Speech to text | Anyone with recordings | A transcript, with the language detected for you
Text to speech | Creators | Audio as mp3, wav, flac or opus
Voice cloning | Creators | A reusable Voice ID made from your voice sample

How to set it up: The hosted page runs once the key is set. A local copy needs these things first.

• Python 3.11 on Windows, macOS or Linux.
• pip, which installs the packages.
• System packages from packages.txt, which Hugging Face installs for you.
• A paid Mistral API key with billing active.

1. Install the huggingface_hub package with pip.
2. Download the Space's files into a folder with the hf command.
3. Move into that folder.
4. Install Gradio 6.12.0, then the packages listed in requirements.txt.
5. Set your Mistral key as MISTRAL_API_KEY, then start the app with Python.

pip install -U huggingface_hub
hf download codexxx/voxtral-studio --repo-type space --local-dir voxtral-studio
cd voxtral-studio
pip install gradio==6.12.0
pip install -r requirements.txt
python app.py

On Hugging Face, add the key in the Space settings under secrets, with this name and value.

MISTRAL_API_KEY = your-key-here

How to use it: Open the app and pick a feature. For speech to text, give it a recording. For text to speech, type your text and pick an output format. For voice cloning, upload a voice sample and keep the Voice ID you get back.

Tool | What it does
Speech to text | Transcribes audio in 13 languages with automatic detection
Text to speech | Makes natural speech as mp3, wav, flac or opus
Voice cloning | Turns a voice sample into a reusable Voice ID

Example: You upload a short interview recording and run speech to text. The app detects the language and returns a transcript. You then paste a paragraph into text to speech and choose mp3, and it returns the spoken audio.

For newcomers and experts: If you are new, get a Mistral API key first, because nothing works without it. Add it as a secret, then try speech to text with a short clip.

If you know the field, the app sends your audio and text to Mistral's API with your key. It uses voxtral-mini-latest for transcription and voxtral-mini-tts-2603 for speech and voice cloning. The code sits in app.py and core.py, with system packages in packages.txt.

You add your Mistral key as a secret
You give audio, text or a voice sample
The app calls the Voxtral model through the API
You get text, audio or a Voice ID back

Good to know: Voice cloning gives you a Voice ID to reuse, so keep it somewhere safe.

• License: apache-2
• Cost to run: a paid Mistral API key with billing active, since there is no free tier for the text-to-speech models
• Data it reaches: your audio, text and voice samples go to the Mistral API
• Not for: use without a Mistral key

On Hugging Face, 13 November 2026

• Likes: 1
• Built with: gradio
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/codexxx/voxtral-studio
• Files: https://huggingface.co/spaces/codexxx/voxtral-studio/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.