Pixelle-Video Releases

This release focuses on resolving SDP parser issues and improving WebRTC compatibility.

Jun 4, 2026

rvoip 0.2.0 is a beta candidate release for the SIP series crates, introducing a refactored developer API for the rvoip-sip library to enhance simplicity. The release prioritizes filling feature gaps, interoperability improvements, and performance enhancements.

Jun 4, 2026

NVIDIA launches Cosmos 3, a next-generation open omnimodal world foundation model for Physical AI, unifying text, images, video, audio, and actions to support robotics and simulation.

Jun 1, 2026

This release features a massive update across the stack focusing on critical bug fixes, performance improvements, and the integration of new models and retrievers. Key additions include new image generation providers, cost tracking capabilities, and enhanced retrieval tools to support more complex research workflows.

May 28, 2026

This release introduces support for converting Lottie stickers to animated WebP and adds an environment variable for controlling status broadcast batch sizes. It addresses critical stability issues, including fixes for sessions stuck due to buffer limits and video format conversion errors. The update also enhances the API and dashboard with new pagination and disappearing message settings.

May 26, 2026

This release focuses on enhancing text path rendering capabilities by adding support for Right-to-Left (RTL) layouts. It also includes maintenance updates to improve the CI pipeline, documentation, and test suite stability.

Apr 30, 2026

This release focuses on resolving various bugs related to web scraping, data parsing, and streaming, while introducing new LLM provider support and a social media retriever. It includes updates to dependencies and fixes for missing imports in websocket managers and PDF reports.

Apr 16, 2026

A maintenance release focusing on Safari browser compatibility, specifically fixing rendering issues related to text, shadows, and opacity.

Apr 10, 2026

A maintenance release addressing execution order issues and updating distribution links.

Apr 10, 2026

MNN 3.5.0 focuses on three core themes: multi-backend LLM inference expansion, high-performance quantization & sampling optimization, and on-device voice interaction enhancement.

Apr 7, 2026

A maintenance release focused on stabilizing the application by fixing a crash that occurred during drag-and-drop operations involving transformers.

Mar 16, 2026

A minor patch release aimed at correcting structural issues within the application and ensuring the build configuration is up to date.

Mar 13, 2026

A comprehensive release featuring major performance improvements, bug fixes, and new frontend capabilities for the FastAPI interface.

Mar 13, 2026

MNN 3.4.1 introduces Qwen3.5 model support with Linear Attention operators across multiple backends, along with significant improvements in LLM resource management and memory security.

Mar 5, 2026

This release focuses on performance optimizations, bug fixes, and new provider support for the GPT Researcher project. It includes smart context compression, URL deduplication, and Docker compatibility improvements.

Mar 1, 2026

MNN 3.4.0 focuses on deepening GPU/QNN backend capabilities, optimizing attention computation for long-context scenarios, and improving GPU runtime stability on iOS platforms.

Feb 7, 2026

This release updates the Surya library to version 0.17.1 to address upstream issues and includes minor updates to documentation, licensing, and development workflows.

Jan 31, 2026

This release focuses on improving LLM configuration and stability by introducing automatic model loading and connection testing. It also addresses a bug in digital voice-over text transmission and adds a reminder for SelfHost workflow validation.

Jan 27, 2026

This update introduces the 'Action Transfer' module for video reference processing and expands local workflow support for digital humans and image-to-video generation. It also locks the edge-tts version to ensure audio generation stability.

Jan 26, 2026

A major feature release that introduces digital human broadcasting pipelines, audio-visual synchronization workflows, and extensive multilingual TTS support. It also adds hardware support for high-end RunningHub machines.

Jan 14, 2026