About Prompsit.
Expand your language horizons.

Our Mission

Since 2006, Prompsit Language Engineering has transformed linguistic data into technology for business: quality datasets, domain-specific machine translation, and reliable, evaluable NLP pipelines.

  • Data curation: collection, cleaning, alignment, and enrichment of multilingual corpora.
  • Modelling: training and fine-tuning of MT/LLM for regulated domains and enterprise terminology.
  • Delivery: secure on-prem or cloud deployments, with reproducible metrics and end-to-end traceability.

How We Work

  • Traceable pipelines, continuous QA, and metrics-driven decisions.
  • Commitment to low-resource languages and preservation of linguistic diversity.
  • Privacy-by-design approach: data and models under client control (on-prem or private VPC).

Open-Source

We maintain an open-source ecosystem of tools for extraction, cleaning, alignment, evaluation, and training. Our solutions are adopted by the community, companies and public administrations across Europe and serve as the foundation for numerous projects.

European R&D and Digital Sovereignty

We participate in R&D initiatives such as OpenEuroLLM and HPLT (High-Performance Language Technologies) to promote open, transparent models and strengthen European digital sovereignty.

Roots

Spin-off from the Transducens group (University of Alicante). Based at the Science Park of the Miguel Hernández University (UMH), in Elche.

Principles

  • Technological excellence: scientific rigor oriented toward results.
  • Transparency: open software, traceability, and honest evaluation.
  • Collaboration: joint work with clients, universities, and the public sector.

Key Figures

  • 7.5 raw PB of multilingual datasets.
  • 200+ languages covered.
  • 250+ trained/fine-tuned models (MT and language).
  • 15+ open-source tools maintained by Prompsit.
  • 20 years in the market.

FAQ

  • What is Prompsit Language Engineering?

    Spanish company (2006), spin-off from Transducens (University of Alicante), specialised in language engineering and AI for NLP.

  • What does Prompsit do in practice?

    We curate and align multilingual data, train and fine-tune domain-specific translation/LLM models, and deploy secure solutions on-prem or in the cloud with reproducible metrics.

  • Do you work with low-resource languages?

    Yes, it’s a priority: we collect and enrich corpora for low-resource languages and preserve linguistic diversity.

  • What sets you apart from generic MT/LLM providers?

    Data and domain: traceable pipelines, clear metrics, enterprise terminology, and a strong commitment to open source and digital sovereignty.

  • Can you handle sensitive data and compliance?

    Yes. Privacy-by-design approach and deployments on-prem or in the client’s VPC, with full control of data and models.

  • What size and variety of data do you manage?

    7.5 raw PB in 200+ languages, including regulated domains (legal, healthcare, financial, etc.).

  • How many models and tools do you maintain?

    Over 250 trained/fine-tuned models and 15+ open-source tools for extraction, cleaning, alignment, evaluation, and training.

  • Which European initiatives do you participate in?

    OpenEuroLLM and HPLT.

  • How does a project with Prompsit start?
    1. Discovery and sampling
      1. Curation/alignment plan and metrics
      1. Pilot with evaluation
      1. Secure deployment and transfer.
  • Where are you based and who do you collaborate with?

    In Elche (UMH Science Park); collaboration with universities, companies, and the public sector in Europe.

Let's build your next AI product together

Reach out and our innovation team will get in touch within 24 hours.

Contact us
About Prompsit | 20 Years of Language Technology Innovation | Prompsit