What's new in Kuda Lingo

Translation & Security Update

This update introduces several improvements to the UX and performance.

• Added
  LlamaSharp for real-time translation

  Added LlamaSharp and asynchronous translation for improved speed and flexibility.

• Added
  Translation inference settings

  Added settings for translation inference.

• Fixed
  Security vulnerability in third-party dependency (CVE-2025-6965)

  Fixed a security vulnerability by removing LM-Kit.NET dependency.

• Improved
  Translation inference speed

  Improved translation speed by making it fully asynchronous.

• Improved
  Reduced app package size by removing LM-Kit.NET dependency

  Replaced the LLM inference back-end with LlamaSharp and removed LM-Kit.NET dependency.

Bug fixes & Optimization Update

This update introduces several improvements to the UX and performance.

• Added
  Third-party attributions

  Added credits for third-party open-source projects.

• Fixed
  "Detect" language option when translating

  Fixed the unreliable automatic language detection from speech when selected.

• Fixed
  Language options when selecting models

  Fixed unreliable language options when selecting translation models.

• Fixed
  Quantization option when selecting transcription models

  Fixed unreliale quantization options when selecting transcription models.

• Improved
  Processing time for segment counting

  Improved the processing time when segments are counted for a 1-2s reduction.

• Improved
  Translation time

  Improved translation time by making it run asynchronously.

Multilingual Translation Update

This update focuses on additions and improvements to the UI and user-experience.

• Added
  Voice Activity Detection

  Added a VAD pre-pass to the transcription process to filter audio.

• Added
  Multilingual translation models

  Translations are now available in multiple languages.

• Added
  Small model

  Added small model to the Basic (free) tier.

• Added
  Translation window

  A new translation window now shows along with the transcription.

UI & UX Update

This update focuses on additions and improvements to the UI and user-experience.

• Added
  New UI window for models

  Added a new UI window to manage downloaded models and changed the flow of model selection when starting a transcription.

• Added
  Model selection for Basic tier

  Users are now able to select models in the Basic (free) tier.

• Added
  Feedback buttons

  Added buttons to give feedback or report AI hallucinations.

• Improved
  More transcription details

  The details panel now has more information about the transcription process, also known as inference.

• Improved
  Better hardware vendor details

  The details panel now shows which CPU vendor, such as Intel or AMD, is being used for the transcription process.

Initial Preview Release

The foundational proof-of-concept launch bringing private, fully offline AI transcription & translation capabilities onto the Windows desktop environment. Initially only with live transcription in 95 languages and translation to English.