HiTZ

What HiTZ does

HiTZ (Basque Center for Language Technology / HiTZ Zentroa) is an academic research center hosted at the University of the Basque Country (UPV/EHU) in Donostia-San Sebasti\u00e1n, Spain, with a secondary office in Bilbao. Formally established as HiTZ in 2019, it consolidates two long-standing research groups, Ixa (founded 1988) and Aholab (founded 1995), that together represent the principal driving forces behind Language Technologies in the Basque Country. The center's mission is research, training, technological transfer, and innovation in AI focused on language and speech, with particular emphasis on Basque as a low-resource language.

The center develops and maintains a broad portfolio of language technology assets: reference corpora (EPEC, 300,000 words; Garaterm, ~18 million words), lexical databases (Basque WordNet, EDBL), speech synthesis and recognition systems (AhoSpeaker, Ahosyn, Iparrahotsa, ZureTTS), NLP tools (leXkit, Arikiturri), evaluation benchmarks (BasqueGLUE), and the Latxa open large language model for Basque with an associated chatbot and translator. Underlying technology spans NLP, machine translation, text analysis, speech synthesis, and multimodal LLMs, running on dedicated GPU infrastructure (28 A100 80Gb plus 8 H100 GPUs) supplemented by EuroHPC allocations.

HiTZ operates as a non-commercial academic entity with a 51-100 person multidisciplinary team. Revenue is generated through research grants (Spanish MICIU/AEI, EU NextGenerationEU/PRTR, ERDF, EuroHPC, CLARIN, CLARIAH), contracted research with public institutions (Osakidetza Basque health system, Galdakao-Usansolo Hospital) and enterprise clients (Minsait in legal NLU, Hispavista in medical QA), and technology transfer including the ENIA/SEDIA-backed Chair of AI and Language Technologies inaugurated in September 2025. The center is also a research-community node, publishing regularly in NeurIPS, ICLR, AAAI, and ACL, training doctoral and master's researchers, and participating in European research infrastructure networks.

Name

HiTZ

Legal name

HiTZ Zentroa - Basque Center for Language Technology

Website

https://hitz.eus

Company type

Private

Founded year

1988

Operating status

Operating

Headcount range

51–100 employees

Short description

HiTZ is an academic research center at the University of the Basque Country specializing in AI and language technologies, primarily for Basque and other low-resource languages. It develops corpora, speech systems, and the Latxa LLM, serving healthcare, legal, and academic clients through research grants and contracts.

Ownership category

akta.pro rank

Product category

Language Technology / Natural Language Processing

NAICS

Research and Development in the Social Sciences and Humanities (541720), Computer Systems Design and Related Services (54151), Scientific Research and Development Services (5417)

SIC

Services-Computer Programming, Data Processing, Etc. (7370), Services-Engineering, Accounting, Research, Management (8700)

akta.pro industry

PRIMARYMPAHAFAD

Translation & Transcreation Services (Script/Dialogue Adaptation)

Media, Entertainment & Sports/Media Technology & Services (Post-Production, VFX, Animation, Streaming Infrastructure)/Localization & Versioning (Subtitling, Dubbing, QC)

Keywords

• Natural language processing
• Language technology research
• Machine translation systems
• Speech synthesis tools
• Basque language AI

Headquarters

HQ city

Donostia-san Sebastián

HQ country

Spain

HQ region

Europe

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Infrastructure, Operations, Marketing or Sales

Revenue model

1. Research Contracts: Contracted research projects with hospitals (Osakidetza), government institutions, and companies (Minsait) for NLP solutions in medical and legal domains.
2. Research Grants: Funding from public sources including Spanish government (MICIU/AEI), European Union (NextGenerationEU/PRTR, ERDF, EuroHPC), and European research infrastructure projects (CLARIN, CLARIAH).
3. Technology Transfer: Transfer of developed NLP tools, language resources, and technologies to industry and public sector, including the Chair of AI and Language Technologies.

Go-to-market motion2 records

Distribution channels1 record

Marketing channels6 records

Core offering

HiTZ is a research center that develops and provides language technology tools, language resources (corpora, lexical databases), NLP systems, speech synthesis/recognition systems, machine translation, and the Latxa large language model, primarily for Basque and other languages. The center delivers these capabilities through research contracts with healthcare, legal, and government organizations, and openly releases tools and datasets for the research community.

Product overview

HiTZ (Basque Center for Language Technology) is a research center at the University of the Basque Country comprising the Ixa and Aholab research groups. The center offers a comprehensive portfolio of language technology products and resources including corpora (EPEC Corpus with 300,000 words), lexical databases (Basque WordNet, EDBL), speech synthesis systems (AhoSpeaker, Ahosyn, Iparrahotsa), NLP tools (leXkit dictionary CMS, Arikiturri question generator), evaluation benchmarks (BasqueGLUE), and the Latxa open language model for Basque. The center focuses on natural language processing, machine translation, speech processing, and AI applications across medical, legal, and educational domains.

Differentiator

Problem solved

Functional benefit

Brands

• Latxa: An open language model and evaluation suite for Basque
• BasqueGLUE
• HiTZ Chair of Artificial Intelligence and Language Technology

Products and services

• EPEC Corpus The EPEC (Euskararen Prozesamendurako Erreferentziazko Corpusa) is the reference corpus for Basque language processing, containing 300,000 words tagged at morphological, syntactic, and semantic levels. It is used by NLP researchers and developers working on Basque language technology.
• Basque WordNet Basque WordNet (Euskal WordNet) is a lexical database for Basque constructed with the expand approach of the English WordNet, providing semantic relationships between Basque words. It is used by NLP systems, researchers, and developers working on Basque semantic processing.
• leXkit leXkit is a generic XML-based Dictionary CMS used for managing and publishing dictionary content in structured XML format. It is used by lexicographers, researchers, and language resource developers.
• Medical Entity Detection System System to detect Spanish medical entities (drugs, disorders, body-part names) in the medical domain, used for clinical NLP applications in healthcare settings such as adverse drug reaction detection and clinical named entity recognition.
• Itzulbide Itzulbide is software for automatic translation of clinical texts from Basque to Spanish, supporting healthcare documentation in Basque language for hospitals and clinical staff.
• ZureTTS ZureTTS is a public voice bank providing open-source text-to-speech resources for Basque language, used by researchers and developers building Basque TTS applications.
• Latxa
• BasqueGLUE
• Arikiturri
• AhoSpeaker
• Ahosyn

Quantifiable outcome

• Machine translation systems developed between Basque and Spanish for clinical domain
• +3 more outcomes

Named customers4 records

Segments4 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability8 records

Feature14 records

Partnerships

Seven partnerships are on record, tiered core, major and minor.

• Galdakao-Usansolo HospitalcoreStrategic or Co-development PartnerStarted collaboration in 2010 with the aim of improving encoding of health records with International Classification of Diseases (ICD). This partnership marked the beginning of HiTZ's health domain work and benefited patient care through better clinical documentation.
• Osakidetza (Basque Sanitary System)coreStrategic or Co-development PartnerMultiple hospitals within the Basque Sanitary System collaborated on OSAKU, DETEAMI, EXTRECM, and PROSAMED research projects. HiTZ received a contract to automatically translate ICD-10 to Basque for health records.
• MinsaitmajorStrategic or Co-development PartnerHiTZ was contracted to collaborate in Prótagoras project for building NLU system for Spanish in legal domain. Focused on notarial deed information extraction without using labeled data.
• HispavistaminorStrategic or Co-development PartnerCollaboration on data privacy in AI for health applications - developing QA system to extract specific information from medical reports for better decision making.
• ENIA/SEDIAcoreStrategic or Co-development PartnerSEDIA (State Secretariat for Digital Transformation) supports the Chair of AI and Language Technologies at HiTZ under the ENIA (National Artificial Intelligence Strategy) framework.
• CLARIN-KcoreStrategic or Co-development PartnerEuropean research infrastructure for language technology and digital humanities. HiTZ collaborates expanding knowledge acquired in CLARIN-K and Ixa Group for Basque language studies.
• CLARIAH-EUScoreStrategic or Co-development PartnerCross-border CLARIAH node for Basque language and culture. Strategic network for digital humanities and social sciences research integrating Basque researchers into European research infrastructures.

Scale indicators10 records

Recent moves6 records

Expansion highlights6 records

Direct peers

• PRHLT (Pattern Recognition and Human Language Technologies): PRHLT is the NLP and speech research center at Universitat Polit\u00e8cnica de Val\u00e8ncia, working on MT, speech, and multimodal language processing. It is a Spanish university-based language-tech research group with similar scientific focus and academic-center structure.
• Vicomtech: Vicomtech is an applied AI research center in San Sebastian (Basque Country) operating very close to HiTZ, with overlapping language technology, speech, and computer vision work for industry. It is the most directly comparable applied-research peer in the same ecosystem and geography.
• DFKI (German Research Center for Artificial Intelligence): DFKI is Germany's largest AI research center, with major speech and language technology groups. It mirrors HiTZ as a publicly funded, university-affiliated research center with industry contracts, multilingual NLP focus, and a strong publication record.
• Fondazione Bruno Kessler (FBK) \u2013 HLTC: FBK's Human Language Technology center is one of Europe's leading public research units in NLP, MT, and speech. It is a strong peer because of its size, public funding model, multilingual focus, and mix of academic and industry-funded research.
• University of Helsinki \u2013 Helsinki-NLP: Helsinki-NLP is a university research group known for large open multilingual NLP models (e.g., the OPUS project) and for low-resource and massively multilingual MT. It is a direct scientific peer in open multilingual NLP and language-resource production.
• CiTIUS (Centro Singular de Investigaci\u00f3n en Tecnolox\u00edas Intelixentes): CiTIUS is a Galician university AI research center with a strong NLP and language technology group (including Galician-language work), publishing at similar venues and operating under a similar university-based research-center model.

Broad incumbents

• Barcelona Supercomputing Center (BSC): BSC is Spain's national supercomputing center with a large NLP and AI program and significant HPC resources, partially overlapping with HiTZ in language technologies and EuroHPC compute access. It is a broader incumbent rather than a niche specialist in low-resource languages.
• INRIA: INRIA is France's national computer science and AI research institute, with multiple NLP and speech groups (e.g., ALMAnaCH). It is a broader incumbent with overlapping scientific interests and a similar publicly funded research-center model.

Others

• Hugging Face: Hugging Face is the dominant open platform for hosting and distributing language models and datasets, including low-resource language models like Latxa. It is an ecosystem / distribution peer rather than a research competitor, but it strongly influences how HiTZ's open assets are consumed.
• CLARIN ERIC: CLARIN is the European research infrastructure for language resources and technology, of which HiTZ is a member via CLARIN-K. It is an ecosystem / infrastructure peer rather than a direct competitor, but it shapes the funding and standards HiTZ operates within.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks7 records

Key highlights7 records

Customer concentration

• Xx.com/Hitz_zentroa
• YouTubeyoutube.com/@HiTZzentroa/playlists
• LinkedInlinkedin.com/company/hitz-zentroa

Revenue estimate

Valuation estimate

Number of profiles

Profiles4 records

Subsidiaries2 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about HiTZ