sam-audio Releases

A major architectural update introducing PBL v2, re-licensing the project to MIT, and restructuring the codebase into modular npm packages. It adds new AI models and editing tools.

Jun 29, 2026

Introduces new recognizers for Germany, Sweden, and Canada, along with a 'Slim' NLP engine and improvements to existing recognizers like the UK driving license.

Jun 28, 2026

Introduces a Pro Mode editor for slides and outlines, improves offline capabilities, and expands the ecosystem with new AI models, search providers, and speech-to-text support.

Jun 2, 2026

This update adds support for Gemini embedding dimensions in Catsu wrappers, introduces PyEmscripten wheels for WebAssembly support, and fixes JSON export encoding to handle non-ASCII text correctly.

Jun 1, 2026

NVIDIA releases Cosmos 3, a next-generation open omnimodal world foundation model for Physical AI. This release unifies language, images, video, audio, and actions within a single architecture to enable developers to build agents for understanding, reasoning, simulating, and acting in the physical world.

Jun 1, 2026

This release introduces MistralOCR integration for extracting text from images and includes a fix for initial CodeChunker language downloads.

May 19, 2026

This release focuses on a significant performance overhaul of the CodeChunker, reducing index time by nearly half and improving latency across the board based on benchmark comparisons.

May 13, 2026

Enhances model inference capabilities with GPU support, introduces new NER recognizers (HuggingFace, Medical), and adds recognizers for UK and Nigeria infrastructure.

Mar 18, 2026

Adds new recognizers for Thailand and India, enhances infrastructure with Python 3.13 support and Docker security improvements, and integrates language models.

Feb 12, 2026

Oct 21, 2025

Introduces advanced healthcare integration with Azure Health Data Services, Korean Resident Registration Number recognition, and significant security/CI improvements including OIDC for PyPI.

Sep 9, 2025

This release introduces GPT-5 support and enhances node reliability by allowing the Delegate Function Call node to fall back to external functions. It also optimizes recording sizes and updates various model providers.

Aug 8, 2025

Minimal release notes: "## New Features * Optimized serialized recordings - they should be much smaller now"

Jun 30, 2025

Jun 2, 2025

This release updates supported AI models and addresses critical bugs affecting the AI Graph Creator and interpolation syntax handling.

May 29, 2025

A major feature release introducing Model Context Protocol (MCP) support, Project References, and an AI-powered Graph Creator, alongside expanded AI model capabilities.

May 28, 2025

Sep 2, 2024