
Institute · Barcelona, Spain
Projecte Aina
A Catalan government initiative led by the Barcelona Supercomputing Center that builds open Catalan language models, speech resources and corpora.
Barcelona, Spain
- Founded
- 2021
- Type
- Institute
- Based in
- Barcelona, Spain
- Key papers
- 1
- Featured projects
- 8
About
Projecte Aina is a Catalan government initiative created by the Generalitat de Catalunya's digital policies department to ensure that Catalan is well supported in language technology. The Barcelona Supercomputing Center (BSC) carries out the technical work, with its Language Technologies Unit credited as author of the models.
The project has built a large Catalan text corpus, started a Catalan speech corpus and published models and datasets on Hugging Face under permissive licenses. Its Salamandra family of 2B, 7B and 40B models, released under Apache 2.0, was trained on the MareNostrum 5 supercomputer with Aina funding.
Resources are intended for reuse by companies, public bodies and researchers, and the project also runs community programs such as the Aina Challenge.
Mission and ambitionsAina aims to ensure that technology understands and speaks Catalan, with open resources that others can reuse.
Known for
- 01
Catalan language corpora
- 02
Open Catalan models
- 03
Salamandra models on MareNostrum 5
- 04
Government-backed language technology
Milestones
- 2022
Project begins
BSC starts building Catalan corpora and models for the Generalitat de Catalunya.
- 2023
Catalan models on Hugging Face
Open Catalan language, speech and translation models and datasets are published under the projecte-aina organization.
- 2024
Salamandra
Released 2B, 7B and 40B multilingual models under Apache 2.0, with Catalan upsampled in the training data.
- 2025
Aina Challenge and territorial intelligence programs
Programs run in parallel with the model work to encourage reuse of the resources.
Open source
Catalan language models and resources from the Aina project (27 stars)
Audio transcription tool with offline ASR support, a fork (26 stars)
REST API for synthesising Catalan speech (17 stars)
Pre-production spaCy releases for Catalan (13 stars)
Catalan named entity recognition model.
Catalan-focused language model.
- projecte-aina/CATalogdataset
Catalan text corpus.
Catalan text-to-speech demo.
Research papers
Notable releases
Openness
Models, datasets and code are public on Hugging Face and GitHub.
People
- Barcelona Supercomputing CenterLead institution
- Marta VillegasAuthor of an article on AINA
Listed from the lab's own pages and public profiles.
Contact and community
The Aina site has a contact page for the initiative.
Sources
Compiled from the lab's public pages, Hugging Face and GitHub. Last checked 2026-11-17. Something wrong or out of date? Tell us on Discord.



