Updates

2025-08-01 - Context-Aware Interpreting

We now support contextual profiles for each device (tablet), allowing the interpreting system to better understand the environment in which it is used. This means translations can be adapted to the specific needs of different sectors, such as:

  • Education – classroom interaction and student–teacher communication.

  • Healthcare – patient dialogues, medical terms, and everyday explanations.

  • Social welfare – family support, case handling, and community services.

  • Migration and integration – registration, orientation, and everyday questions.

  • Justice system – legal terminology, court sessions, and police procedures.

With context awareness, translations become more accurate, relevant, and professional – ensuring that communication feels natural and appropriate for each situation.

2025-07-25 - Improved Language and Dialect Support

We have introduced enhanced system prompts to improve accuracy and naturalness in several key languages and dialects. These updates ensure that translations sound more authentic and better reflect everyday speech, rather than defaulting to overly formal language.

The improvements cover:

  • Arabic dialects: Gulf, Iraqi, Levantine (North, Central, South), Egyptian, Maghrebi, Sudanese, and Modern Standard Arabic (MSA).

  • Other major languages: Tigrinya, Somali, Persian (Farsi), Dari, Kurdish (Sorani and Kurmanji).

With these updates, the system:

  • Produces output that is closer to real spoken language.

  • Differentiates clearly between dialects (e.g. Gulf Arabic vs. Iraqi Arabic).

  • Handles official terms consistently (e.g. personnummer, BankID, Försäkringskassan).

  • Reduces errors from mixing in words from the wrong dialect or from Modern Standard Arabic.

These improvements are part of our ongoing effort to provide more natural, accurate, and culturally appropriate interpreting and translation for users in Sweden and beyond.

2025-07-10 - Translation & Audio Quality Evaluation

We have introduced a new evaluation system that allows each interpreting session to be reviewed. After a session, users can now provide feedback on both:

  • Text quality – clarity, accuracy, and appropriateness of the translation.

  • Audio quality – pronunciation, naturalness, and technical clarity of the voice output.

This feedback will be used to continuously improve both text translations and speech output, ensuring that our interpreting system becomes more accurate, reliable, and user-friendly over time.