MMAudio Releases

This release focuses on making inbound voice AI production-ready by introducing distributed SIP registration, multi-server ownership, and enhanced failover mechanisms. It also introduces multi-target SIP transfer support and defines explicit post-transfer behaviors.

Apr 28, 2026

This release introduces async observability writes to reduce latency, adds support for vLLM's reasoning field and TypeScript evaluators, and enhances UI capabilities with aggregated variant usage data and cost tracking in traces.

Apr 24, 2026

This release primarily focuses on fixing bugs related to OAuth2 error handling, middleware issuer checks, and server health checks. It includes improvements to error reporting and environment variable writing. Additionally, this version welcomes two new contributors to the project.

Apr 12, 2026

This release addresses a specific bug in the Duo integration where incorrect response codes were checked for successful API calls. It provides updated Docker images for users to deploy the latest version.

Apr 10, 2026

This release focuses on stabilizing authentication and authorization flows, fixing cache misses, LDAPv3 compatibility, and domain matching. It introduces configurable rate limits for OIDC and welcomes several new contributors to the project.

Apr 9, 2026

Rapida v2.1.0 introduces built-in observability with granular per-stage latency tracking, a major streaming pipeline architecture rewrite, and a comprehensive UX overhaul. The release also adds JSON-driven provider configuration and inline noise reduction capabilities.

Apr 1, 2026

This release introduces pluggable Voice Activity Detection and End-of-Speech engines, improves audio handling with a heartbeat mechanism, and upgrades infrastructure to Go 1.25.8 while adding comprehensive test coverage.

Mar 17, 2026

This release focuses on security fixes and resolves various bugs related to authentication, OIDC, configuration, and web UI functionality.

Mar 14, 2026

This release adds new capabilities to the Katana tool, including URL path pattern deduplication, support for captcha detection and solving in headless mode, and knowledge base classification using DIT.

Mar 10, 2026

This release introduces a new Fluxer notification provider and updates the process title for Uptime Kuma. It also includes bug fixes for UI styling, return types, and a security patch to address Server-side Template Injection (SSTI).

Mar 10, 2026

Feb 7, 2026

This release introduces headless mode support for the crawler and resolves a concurrency issue within the hybrid crawler.

Jan 6, 2026

This release focuses on fixing ESM import path errors and resolving the visibility of default Vimeo captions.

Jan 3, 2026

This release updates the inference pipeline and models to support unlimited audio length, alongside updates to the model cards.

Dec 22, 2025

Dec 12, 2025

This release focuses on enhancing CLI usability and utility functions, alongside fixing compatibility issues with the latest PyTorch Lightning version. It introduces a new revision option for commands and a safer calibration transform method. Additionally, it updates the pyannote-database dependency to version 6.1+.

Dec 7, 2025

This release introduces new filtering capabilities for CLI output and resolves several bugs related to option validation, JavaScript event detection, and scope behavior. It also includes improvements to logging depth, error handling, and program exit status.

Dec 1, 2025

A maintenance release addressing dependency stability, updating the pyannoteAI wrapper, and adding convenience methods for pipeline execution.

Nov 19, 2025

Oct 10, 2025

A major version bump introducing VBx clustering, exclusive speaker diarization, faster training, and pyannoteAI premium integration, while dropping support for older Python versions and audio backends.

Sep 29, 2025