unmute Releases

This release focuses on usability, runtime stability, and streaming improvements, introducing a validation command for training manifests and safer LoRA loading mechanisms.

May 11, 2026

Fixes version detection issues where `dx --version` used an outdated git SHA and `dioxus-cli` installed via cargo had no version, preventing self-update from functioning properly.

May 8, 2026

Fixes small bugs discovered during v0.8 development, including macOS signing issues, TOCTOU race conditions in wasm hotpatching, and an iframe fix.

May 7, 2026

This release addresses a regression in JavaScript bundling where snippets were incorrectly always bundled as ESM. It now automatically detects the module type (ESM, CJS, UMD) and runs esbuild with the appropriate flags to ensure correct bundling behavior.

May 1, 2026

This release serves as the final release of the 0.7 series, fixing numerous bugs and adding features such as shell completions, a faster dev server, and Windows app icon bundling.

Apr 22, 2026

Apr 21, 2026

A comprehensive update to Umami analytics featuring Boards, Session Replay, Web Vitals tracking, and a redesigned share page.

Apr 16, 2026

This is a minor bug fix release for version 2.0.2.

Apr 8, 2026

This release removes the auto-trim feature for reference audio.

Apr 8, 2026

The major version 2.0.0 release of VoxCPM featuring code fixes, performance optimizations, and dependency updates.

Apr 6, 2026

This release focuses on hyper-parameter precision and accurate model support across providers. It adds support for new frontier models like Anthropic Opus 4.6 and GPT-5.4, alongside a new AWS Bedrock integration. The update emphasizes validated parameters and customizable reasoning controls for various AI providers.

Mar 25, 2026

This maintenance release for Frigate 0.17 includes security fixes and minor changes to improve stability. Notable updates include restricting the raw config endpoint to admin users and fixing cross-camera authentication issues. Additionally, it resolves various UI display conditions and documentation errors.

Mar 22, 2026

This service release focuses on security hardening, interoperability improvements, and bug fixes to create a stable platform. Key additions include easier Ollama configuration via environment variables and improved fallback response handling for caption generation. The update also addresses reliability issues in indexing, Places, folder albums, and metadata parsing.

Mar 5, 2026

This major release introduces significant new features such as classification model training, custom viewer roles, and semantic search triggers. It also enhances hardware support with CUDA Graphs for Nvidia GPUs and OpenVINO for Intel NPUs. However, it includes several breaking changes regarding GenAI configuration restructuring and recording retention policies.

Feb 27, 2026

Feb 10, 2026

This release introduces significant upgrades including native xAI protocol support, support for the latest AI models like Kimi K2.5 and GPT 5.2, and new integrations such as Google Drive and Inworld speech synthesis.

Feb 2, 2026

Jan 29, 2026

This release addresses a critical security vulnerability involving authenticated remote command execution and container escape. It also includes documentation updates and minor improvements to camera configurations and OpenVINO settings.

Jan 29, 2026

A patch release specifically targeting a security vulnerability in Next.js, as detailed in the official security advisory.

Dec 12, 2025

This security release addresses a broken access control vulnerability allowing viewer users to delete admin accounts. It also improves recording playback efficiency and fixes issues with YOLOv9 export scripts and certificate generation.

Dec 6, 2025