SenseVoice Releases

A maintenance release that fixes a port change issue in the development server and updates documentation regarding browser permissions.

Apr 18, 2026

A short release addressing a specific configuration issue for Firefox Manifest V3 extensions to ensure correct action key handling.

Apr 17, 2026

A minor release focusing on correcting a color name usage and reorganizing the documentation structure for WXT Modules.

Apr 16, 2026

An enhancement-heavy release adding support for Firefox data collection permissions and a new strict port option for the dev server, alongside fixes for script injection and Vite 8 compatibility.

Apr 14, 2026

This release focuses on enhancing input flexibility and data processing capabilities. It introduces a Parquet reader, implements NLP streaming, and updates documentation. Security patches for NLTK and improvements to vector store management are also included.

Apr 13, 2026

A minor patch release dedicated to resolving a specific security vulnerability (CVE-2025-14009) found in the NLTK library.

Mar 27, 2026

This release features a major architectural overhaul by replacing the tokenizer code with skeletoken and refactoring the persistence layer. It also deprecates Python 3.9 support and updates benchmarks to MTEB V2.

Mar 26, 2026

This patch release prioritizes security and stability by pinning the LiteLLM dependency to address a compromise. It also reconfigures the vector store size based on the embedding model to optimize performance.

Mar 24, 2026

This release improves data integrity by eliminating phantom entity relationships and introduces NLP streaming capabilities.

Mar 6, 2026

This update enhances LLM integration with automatic model loading and connection testing, improves workflow stability with SelfHost reminders, and fixes a bug in digital human voiceover generation.

Jan 27, 2026

The release introduces an action transfer module and local workflow support for digital humans and image-to-video generation, alongside a critical stability fix for TTS services.

Jan 26, 2026

This version adds core video generation pipelines for digital humans and audio-visual synchronization, expands TTS to multiple languages, and optimizes Docker deployment and UI/UX.

Jan 14, 2026

Updates focus on infrastructure flexibility, adding new workflows (Nano Banana), configuring cloud resources (ApiKey, concurrency), and improving LLM parsing and configuration hot-reload capabilities.

Dec 28, 2025

This release improves user support with a built-in FAQ sidebar, introduces a new static excerpt template, and resolves TTS service stability issues.

Dec 10, 2025

Adds significant machine learning model performance improvements including quantization and distillation features, alongside Python 3.13 support and configuration updates.

Oct 5, 2025

This release focuses on stability and performance improvements for the Docker environment, specifically addressing session processing issues and optimizing replay capabilities.

Aug 8, 2025

A comprehensive update involving Kafka configuration, new logging capabilities, and support for remote video inlining, alongside various performance and security fixes.

Jun 23, 2025

This update primarily focuses on UI enhancements, specifically adding Material Design icon support and fixing issues with widgets and keyboard interaction.

Jun 10, 2025

This release introduces smaller tokenizers and supertokenizers, allows explicit validation set passing, and improves distillation workflows.

Jun 3, 2025

This update enhances the custom API widget with new functions like modulo and safeHTML, and adds options for JSON string conversion. It also allows disabling the theme picker and applying CSS based on the selected theme.

May 19, 2025