Transcripción automatica en aragonés

Transcripción automatica en aragonés

Transcribes Aragonese speech to text from an uploaded audio file or a microphone recording of up to 60 seconds. A demo of an Omnilingual LLM-3B model fine-tuned for Aragonese.

Made by Juanpabl 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: This demo turns Aragonese speech into text. You give it a saved sound file or speak into a microphone, and a transcript comes back. It runs a model that was fine-tuned for Aragonese, and the Space runs on Hugging Face, so you can use it in a browser with nothing to install.

Who it suits: People who work with Aragonese recordings and want a first-draft transcript, and developers who want to see a speech model packaged as a Gradio app.

Industry | Who uses it | How they use it | Example
Language work | Researchers and speakers | Get a draft transcript of an Aragonese recording | Upload a short interview clip and read the draft
Education | Teachers and learners | Check how spoken Aragonese turns into written text | Record a sentence and compare it with the text
Media | Writers and editors | Draft notes from a short spoken clip | Download the transcript as a text file
Software | Developers | Study how the app wraps an Aragonese speech model | Download the Space and read app.py and inference.py

Anyone who needs live captions, or recordings longer than 60 seconds in one go, can skip it.

Where to use it: It runs in these places.

• A web browser, on the Space's page at Hugging Face
• Windows, with Python 3.10
• macOS, with Python 3.10
• Linux, with Python 3.10

Use case | Who | What they get back
Transcribe an audio file | Researchers and editors | A text transcript you can copy
Transcribe a microphone recording | Speakers and learners | A transcript of what you recorded
Save the result | Anyone | The transcript as a downloadable text file

How to set it up: Using the Space online needs only a browser. A copy on your own computer needs the following.

• Python 3.10 and pip
• The huggingface_hub package, which gives you the hf command
• The system packages listed in packages.txt, which the hosted Space installs with apt-get
• Disk space for the repository, which holds 11 files

1. Install the huggingface_hub package.
2. Download the Space's files into a local folder.
3. Move into that folder.
4. Install Gradio at the version the Space uses, then install the packages in requirements.txt.
5. Start the app.

pip install -U huggingface_hub
hf download Juanpabl/asr-aragones-demo --repo-type space --local-dir asr-aragones-demo
cd asr-aragones-demo
pip install gradio==5.49.1
pip install -r requirements.txt
python app.py

How to use it: Open the Space's page, or run the app on your computer and open the page Gradio shows. Give it a saved sound file or speak into a microphone, then press the Transcribe button.

python app.py

Tool | What it does
Transcribe button | Runs the model on your audio
app.py | The main file, it builds the Gradio page
inference.py | Holds the code that runs the model
packages.txt | Lists the system packages the app needs

Example: You upload a 40 second recording of someone speaking Aragonese and press Transcribe. You get back the text of what was said, which you can copy or download as a text file.

For newcomers and experts: Aragonese is a language of northeastern Spain, and this page writes down what a speaker says. The first small thing to try is the online Space with a short recording. Check the transcript against the audio, because it can contain errors and gaps.

The Space names one model, an Omnilingual LLM-3B fine-tuned for Aragonese on top of Meta's Omnilingual ASR. It handles one recording at a time and accepts clips up to 60 seconds. It is set to run on Hugging Face's zero-a10g hardware, so availability depends on GPU quotas and queues.

You upload or record audio
The Aragonese model reads the audio
It writes the transcript
You copy it or download a text file

Good to know: Only use recordings you are cleared to send to this service.

• License: apache-2
• Cost to run: none named
• Data it reaches: the audio you upload or record, which is processed on Hugging Face infrastructure
• Not for: live captioning, since it handles one recording at a time
• Hardware: the Space is set to zero-a10g, and a busy shared GPU can slow or block a run

On Hugging Face, 13 November 2026

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

Links

• Space: https://huggingface.co/spaces/Juanpabl/asr-aragones-demo
• Files: https://huggingface.co/spaces/Juanpabl/asr-aragones-demo/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.