Prompt-Experts trained in a few weeks claim to automate most business activities.
What's next? Where will you find differentiating added values?


Your cub[AI]x linguistic R&D AI-Experts are True AI Creators (AI software developers),
highly skilled in designing unique technologies for you.

With cutting-edge Open-Source solutions (NMTs, LLMs, etc)
mixed with pure original algorithms and Deep-Learning models
cub[AI]x can provide you with stuning results
protecting your data from Big Tech!

NeuroSpell: Universal Auto-Corrector

Spelling, grammar, transcription, translation, etc

Design your custom voice/text AI-powered process to be run on your own Intranet with cutting-edge technologies:

  • Speech-to-Text including advanced AI-powered voice extraction, noise filtration, and music separation.
  • Automatic spelling and grammar correction combining Deep Learning and Rule-Base systems.
  • Rephraser to get multiple propositions using latest LLMs technologies.
  • Translation with in-domain NMTs specifically trained on your own vocabulary and phrasing.
  • Text-to-Speech with multi-modal LLMs.

Give it a try on our NeuroSpell demonstrator!

Discover why checking is not correcting!

Use case: call center dictation

Live-Chat, customer support, appointment scheduling, etc

Perfect answers in minutes!

Call center employees may not be confident in their writing skills, even when they are very good at their core tasks. This can be very damaging to a company's image.

With our powerful AI-Transcriber, they gain in productivity without any spelling / grammar errors in the text they produce.

They are then secured by our Deep Learning Auto-Corrector when doing adjustments in the obtained text before sending it to the customer.

All the process is deployed On-Premise without any data sent outside. No third-party Big Tech in the loop.

Minimal keystrokes for error-free communication in a totally secured configuration.

KaraLyr: song and video subtitling

Stem separation, word-level synchronisation, etc

KaraLyr's AI-powered technology extracts remixable sub-tracks from any song (vocals, drums, bass, etc).

Our unique high-precision word-level lyric synchronisation ensures you'll never miss a beat with the exact original words!

Also works for subtitling speech in videos.

Free / Open-Source Softwares: our Know-How demonstrators

Documents, music, images, videos, etc

karakok-AI is an Open-Source karaoke Player / Editor with automatic clip creation from any song file using AI-powered vocals and lyrics extraction (Speech-to-Text).

Vocals, Drums, Bass, and Other music sub-tracks can be Live-remixed, for a stuning karaoke experience!

kaiDJ included: a minimalist and easy to use, but yet powerful and versatile DJ Party Player.

Find it on karaok-AI.com or GitHub.

ComPair is an universal document comparator: multi-lingual, cross-lingual, multi-formats.

Compare 2 documents, before/after a modification or before/after a translation, to get a comparison report you can exchange by mail to discuss about the changes with customers, translators, partners or lawyers.

All languages supported in any kind of file formats.

Download and start using it with the Quick Start Guide.

ChatMate is an Open-Source ChatGPT versatile file processing automation (Java application). Use ChatGPT ability to handle complex queries to perform advanced processing on a set of files.

Use-Case: SRT file translation may be complex because of possible bad sentence cut, making each text part loosing its context. ChatGPT, with its ability to handle complex processing, by translating the SRT at once as a whole, can produce much better SRT translations than other available translation tools.

Find it on GitHub.

3Denlive is an Open-Source Stereoscopic photo/video montage software.

Use high-level image and video processing filters to produce stuning stereo effects on your images, slideshows, animations and videos.

Download and start using it from 3Denlive.com or GitHub.

Proprietary Softwares: information on demand

NMT, terminology extractor, semantic processing, etc

Libellex is a powerful and versatile CAT Tool and Translation Workflow with advanced processing capabilities.

  • Alignment of document pairs to create translation memories.
  • Terminology extraction to create accurate bilingual glossaries.
  • NMTs with advanced multi-sources evaluation balancing.
  • Workflow to manage all states of a document in the whole translation process.

XSM (Cross-lingual Semantic Memory) is a unique text database being able to handle millions of mono-lingual and bilingual entries with semantic pairing/matching.

NeuroBiSpell is the ultimate Translation Checker (QA) and Translation Auto-Corrector using Deep Learning.