Andalucía Información visits Myukura: technology born in Jerez to reduce the documentation burden of the consultation

On 28 September, journalist Rocío Alfaro published a report on Myukura in Andalucía Información.

The piece runs from the origin of the project to some of the main technology problems we are trying to solve: speech recognition in Spanish, adapting to different accents, clinical documentation and local processing of information.

It has special value for us that the story is told from Jerez.

Myulab Innovation was born here and, although our technology is built using infrastructure and programmes that connect us with organisations outside Andalusia, we continue to develop the project from our own environment.

The problem that gave rise to Myukura

Myukura works on a question that is simple to state and considerably harder to solve:

can artificial intelligence be used to reduce part of the documentation work of a consultation without taking the professional out of the process?

Our approach combines speech recognition and language processing technologies to transcribe conversations and then help structure the information.

The professional’s review remains a mandatory part of the workflow.

According to the company’s own estimates cited in the report, this type of tool could help reduce the administrative burden associated with clinical work by 35% to 40%.

The project is currently still in development and testing.

When language stops being a detail

One of the topics the report dwells on is language.

Many recent advances in artificial intelligence are initially presented with excellent results on English.

When those systems are moved to other languages and, especially, to specialised vocabulary, behaviour can change.

That is why part of our work focuses specifically on Spanish.

In the words of Nicolás da Fonseca, co-founder of Myulab Innovation:

“We are training the models and taking almost everything over to Spanish.”

Within that problem lies an even more local one: accents.

The Spanish spoken in Andalusia, like any linguistic variety, has its own pronunciations and rhythms that are not always equally represented in the data used to develop the models.

For us, working on that diversity is part of the technical problem.

From a personal experience to a company

The report also covers the origin of the idea.

The project began from an experience of Nicolás in Paris, where the difficulties speech recognition systems had with his French pronunciation raised a broader question about how to build tools able to understand conversations in real contexts.

That initial idea evolved over time into a much more specific problem: clinical documentation.

It was not just about transcribing.

We had to understand what information was relevant, who had said it, how to structure it and how to keep the professional within the verification process.

Cloud and local processing

Another topic covered in the interview is our research on hybrid architectures.

We are currently studying both cloud-based systems and lines of local processing.

The motivation is not to automatically replace one architecture with another.

Each has different advantages, limitations and requirements.

But in scenarios involving especially sensitive information, it is important to investigate how much processing can be done within the device itself or the centre’s own infrastructure.

This line is part of Myukura Edge.

Next steps

The article also covers the roadmap communicated by the company: first advancing on the general medicine and psychology use cases and later studying extending the project to veterinary medicine during 2027.

As with any product in development, these stages depend on technological progress and on the tests we carry out.

The aim is to advance progressively, not to add specialties simply to widen the catalogue.

A growth moment for Myulab Innovation

The report comes after several important pieces of news for the company.

During these weeks, Myulab Innovation has been selected for the 8th MareNostrum Hackathon at the Barcelona Supercomputing Center and for Andalusia’s ScaleUp 2026.

The company also holds ENISA’s emerging company certification and takes part in programmes such as NVIDIA Inception, Google for Startups and Microsoft for Startups.

These are different kinds of support and are not equivalent to clinical validation of the product.

But they do allow something important for a technology startup: access to infrastructure, training, networks and environments in which to test the project.

We thank Rocío Alfaro and Andalucía Información for the space devoted to telling our story and, especially, for their interest in understanding the problem behind the technology.