Skip to content

Free tool

Free Catalan translator.

Coming soon, right inside your WordPress.

Tell me when it's out
0 / 5000 characters

Translation

Your translation will appear here.
Translation options

Translate a whole file

UTF-8 .txt or .po files, up to %mb MB. We translate them in the order they arrive and email them to you.

We will only use your email to send you the translation. More in our privacy policy.

Texts are sent encrypted to our server, translated or checked on the spot and never stored, neither the texts nor the results, unless you tick the box to donate them, anonymised, to a public corpus (CC BY). More in our privacy policy.

No machine is infallible: if the text matters, have a person check it. Found a mistake or have an idea? Tell us.

How the translator works

It uses Aina Translator, the open models from Projecte AINA (Barcelona Supercomputing Center), trained specifically for Catalan.

Eight languages, both ways

Spanish, English, French, Portuguese, Italian, German, Galician and Basque, into or out of Catalan.

Valencian forms

Turn on “Valencian forms” and the translation will use forms such as «eixir», «este» or «servix».

Unknown words

Highlights the words the translator doesn't know, so you can check them before publishing.

Check it with the spell checker

One click sends the translation to the Catalan spell checker for a final pass.

Frequently asked questions

Is the Catalan translator free?
Yes: up to 5,000 characters per translation, no sign-up.
Which languages does it translate?
From Spanish, English, French, Portuguese, Italian, German, Galician and Basque into Catalan, and from Catalan into all of them.
Are the texts I translate stored?
No, unless you tick the box to donate them, anonymised, to a public corpus that helps improve Catalan language models.
Which model does it use?
Aina Translator, from Projecte AINA, under the Apache 2.0 licence. For Valencian forms and unknown words, LanguageTool as well. If the results aren't good enough, our plugin for WordPress, WooCommerce, Shopify, Strapi, Django, Drupal and Laravel Filament has more powerful models: find it at escalf.cat.

What does sharing the text mean?

Large language models (LLMs, like the ones behind AI chatbots) learn from text. There is far less text in Catalan than in English or Spanish, which is why they often write it worse. With the texts you share we build an open dataset to train and evaluate models that write better Catalan.

Before storing it, we remove email addresses, phone numbers, web addresses, ID numbers and IBANs. The text, the result and, if you copy it, your final version will be added to the dataset, which we will publish on Hugging Face under a CC BY licence so that any person, university or company can use it. We don't sell it or use it for anything else.

It is entirely voluntary: if you don't tick the box, the text isn't stored anywhere. You can untick it at any time and write to us to have texts not yet published deleted. If you share it, you help keep Catalan warm!

Hey, how about a little warmth in return?

It helps us improve public Catalan models. We would keep the text after removing email addresses, phone numbers, web addresses, ID numbers and IBANs, and it could become part of a public corpus under a CC BY licence that anyone can use.

Do you want to share this text?