drax Releases

This bug fix release addresses race conditions in S3 snapshots and duplicate document returns in remote federated search, along with several miscellaneous improvements to chat workspace settings and API responses.

Jun 29, 2026

A release opening up A2A agent discovery and a richer skills story, while sharpening security posture and landing a durable run/task control plane.

Jun 26, 2026

This critical security release addresses two vulnerabilities (CVE-2026-57823 and CVE-2026-57824) related to improper authentication that could lead to privilege escalation and information disclosure.

Jun 24, 2026

This release focuses on stabilizing the user interface by fixing critical bugs in the SLU client and correcting tooltip interaction behaviors to ensure proper error reporting and visibility management.

Jun 22, 2026

This release introduces HTML wallpaper support, enhances visual animations and system performance, and resolves layering issues with the dock and toolbar to improve the overall desktop experience.

Jun 22, 2026

The first patch release on v0.8.x focused on stabilizing the multi-agent runtime, featuring cascading config edits, Discord slash commands, and cached-input pricing.

Jun 19, 2026

A major release featuring extensive model support (Gemma 4, DeepSeek-V4), performance optimizations (Model Runner V2, FlashInfer), and architectural improvements like multi-tier cache offloading.

Jun 15, 2026

This update focuses on resource download workflows, added development tools, and fixes critical GUI issues including app menu crashes and settings persistence.

Jun 15, 2026

This release adds an Icon Editor widget, improves settings validation to prevent layout breaks, fixes various GUI behaviors including fullscreen and touch interactions, and resolves issues with Discord presence.

Jun 8, 2026

This release introduces a new OmniTTS backend for vLLM, a comprehensive Web UI for voice management and testing, and an environment variable template for service configuration.

Jun 7, 2026

This release introduces multi-language summary capabilities and adds Qwen 3.5 as a built-in AI model. It also improves desktop app stability by adding tray functionality and preventing multiple instances, alongside various UI and backend fixes.

Jun 5, 2026

A maintenance release focusing on critical bug fixes for browser pools, sitemap discovery, and request handling. It introduces new features like dynamic memory snapshots and abort context helpers to improve reliability and control during crawling operations.

Jun 4, 2026

This release introduces new administrative and avatar generation APIs along with corresponding user interfaces. It features a dedicated admin dashboard for configuration management and an avatar generation page for task submission. Additionally, a main index page has been added to serve as the application's entry point.

May 24, 2026

Updated the session ID data type from integer to string to improve system extensibility and facilitate better integration with frontend APIs.

Apr 11, 2026

Performed a major architectural refactoring to modularize the codebase, introducing a plugin-based architecture for digital human models, TTS, and streaming.

Mar 31, 2026

Expanded functionality to support vision modality for product and diagnosis feedback, alongside a migration to JSON for LLM API caching.

Mar 28, 2026

Adds audio import capabilities, improves transcription quality through parallel resampling, and introduces beta feature gating.

Mar 3, 2026

Fixes Whisper model URLs, Windows audio device selection, and improves model management and API connection testing.

Feb 11, 2026

Introduces AI-powered browser automation and fixes issues related to redirects, persistence, and request handling.

Feb 6, 2026

Updates the default model to GPT-5-mini, introduces empirical validation tools and Jupyter widgets, and adds cost tracking and experimental Ollama support.

Feb 2, 2026