Conteranto

Translation that understands culture, not just words.

Today's translation tools get the words right — but miss tone, politeness, and intent.

Conteranto adapts to cultural context, so your message not only translates… it sounds right to a native speaker.

"Meaning is more than words. Tone is meaning"
Learn More Get in Touch
English
Could you send that report?
→
Japanese
レポートを送っていただけますか？
Formal Polite Indirect

The Problem

The hidden failure of machine translation

Most translation systems are linguistically accurate.

But accuracy is not enough.

A sentence can be grammatically correct and still sound rude, weak, overly formal, or even offensive in another language.

Why?

Because cultures treat tone differently.

  • What sounds neutral in English may sound aggressive in Chinese.
  • A polite Dutch request may become too formal in Japanese.
  • English tolerates directness. Persian prefers softening.
  • Some cultures use passive voice to avoid blame, others see it as unclear.

Current tools assume: one style fits all.

They ignore the pragmatic layer of language — tone, relationship, hierarchy, context.

The result:

  • Misunderstandings
  • Accidental disrespect
  • Loss of trust
  • "Technically correct" but socially wrong
Translation without tone isn't communication. It's noise.

Our Solution

Cultural-context layer for translation

Conteranto addresses this gap by introducing a cultural-context layer for translation powered by large language models and explainable pragmatics.

Instead of producing a single static result, Conteranto generates translations that respect cultural communication norms. The system automatically initializes four key parameters based on the target language and communication context.

01

Automatic Initialization

Smart defaults based on language and context (business email vs. casual chat)

02

User Control

Fine-tune parameters with instant feedback on tone changes

03

Explainability

Live explanations of why adjustments matter

04

Cultural Intelligence

Translations that "sound right" to native speakers

Key Features

Four cultural parameters that make the difference

Politeness

Adjust formality levels from casual to highly respectful, matching cultural expectations

Directness

Control whether communication is explicit or uses subtle, indirect phrasing

Formality

Adapt language register for business, academic, or casual contexts

Attribution

Handle credit and responsibility according to cultural communication norms

Interactive Translation Demo

See how cultural parameters affect translation in real-time

Casual Formal
Indirect Direct
Translation Result:
Could you please send that report when you have a moment?

Philosophy

Proud to share these essays that ignited Conteranto

The ideas behind Conteranto emerged from deep reflection on language, culture, and communication. These essays explore the philosophical foundations of our work:

Prelude to a Thinking Future

Prelude to a Thinking Future

Exploring the foundations of AI consciousness and the future of intelligent systems

 The Luminous Weave

The Luminous Weave: When Two AIs Learn to Think Together

A deep exploration of collaborative intelligence and emergent understanding

 The Intelligence That Refuses to Be One

The Intelligence That Refuses to Be One

Rethinking the nature of intelligence beyond singular, unified minds

Research

Publications on cultural understanding and moral reasoning in AI

Our research explores how large language models understand and process cultural variation, moral reasoning, and cross-cultural communication:

ArXiv 2510.05942

EvalMORAAL: Interpretable Chain-of-Thought and LLM-as-Judge Evaluation for Moral Alignment in Large Language Models

Mohammadi, H., Giachanou, A., & Bagheri, A.

EACL 2026 conference

Under Review 2025
ArXiv 2506.12433

Exploring Cultural Variations in Moral Judgments with Large Language Models

Mohammadi, H., Papadopoulou, E., Meijer, M., & Bagheri, A.

ECIR 2026 conference

Under Review 2025
ArXiv 2507.21319

Do Large Language Models Understand Morality Across Cultures?

Mohammadi, H., Meijer, M., Papadopoulou, E., & Bagheri, A.

2nd Workshop on Language Understanding in the Human-Machine Era, ECAI 2025

Published 2025

Our Team

Complementary expertise across academia and industry

HM

Hadi Mohammadi

PhD Candidate, Utrecht University

Expertise in explainable NLP, large language models, and moral reasoning in AI. Bridges research and implementation with practical AI solutions.

JS

Jeroen Sparla

CEO, AcademicTransfer

Strategic experience in international communication platforms and cross-cultural design. Leads user-centered product development.

AB

Dr. Ayoub Bagheri

Associate Professor, Utrecht University

Leads the Natural Language and Text Processing lab. Specializes in explainable and human-centered NLP with methodological rigor.

Concrete Outcomes

Deliverables and research contributions

Demo Environment

A working prototype built on existing LLM APIs with an intuitive UI, demonstrating the cultural translation capabilities in action.

Unique Dataset

User tone adjustment data providing unprecedented insights into cross-cultural communication preferences across languages.

Research Article

Academic analysis of cultural variance in translation styles, contributing to the field of computational linguistics.

Research Impact

Conteranto will generate a unique dataset on user preferences: how speakers of different languages actually want to communicate. These insights will advance the next generation of culturally adaptive translation algorithms.

Our Mission

Where machine translation once aimed for efficiency, Conteranto aims for empathy and resonance, ensuring people worldwide can communicate with the respect, warmth, and nuance they intend, regardless of language barriers.

"Translation that truly understands people, not just their words."

Get in Touch

Interested in learning more or collaborating?

Email Us

conteranto@academictransfer.nl

GitHub

View Our Code

Research

Utrecht University NLP Lab