list Releases

This release focuses on CLI enhancements for `qmd get`, `multi-get`, and `search` commands, including line-numbering defaults, filesystem path rendering, and new format flags. It also includes fixes for launcher behavior and Darwin Metal crashes.

May 29, 2026

Releases PaddleOCR-VL-1.6, a new document parsing model with improved accuracy and support for tables, Chinese ancient documents, and seals. Also adds official SDKs and multi-page TIFF support.

May 28, 2026

Rewrites the `bin/qmd` launcher script as a Node-based polyglot to resolve Windows global npm installation failures, while maintaining a fallback to Bun for environments lacking Node.js.

May 22, 2026

Updates the release publishing workflow to use GitHub Actions with npm Trusted Publishing/OIDC instead of a long-lived `NPM_TOKEN` secret.

May 20, 2026

Introduces significant UI and process improvements for the platform including card grids, guest space management, process functions, and Korean localization support.

May 10, 2026

Addresses a bug in the sandbox proxy iframe timeout behavior to prevent strict mode rejection errors in the MCP UI client.

May 9, 2026

A minor client update for the MCP-UI-Org project that adds specific props to the AppRenderer component to expose host information and capabilities.

May 1, 2026

This release deeply integrates PaddleOCR with the Hugging Face ecosystem, supports flexible inference engine switching, adds document format conversion capabilities, and releases a browser-based inference SDK.

Apr 21, 2026

This release enhances the Seedance 2 I2V image-to-video nodes by adding richer media input options, including Last Frame, Reference Images, Videos, and Audio, while fixing a bug that previously rejected image-only reference runs.

Apr 20, 2026

A major feature update introducing new onboarding flows, expanded AI model support (Kling, Wan, Seedance), workflow improvements like loop edges and client-side polling, and various UI enhancements.

Apr 19, 2026

A security-focused release addressing critical vulnerabilities like Mass Assignment, IDOR, and Credential Leaks, alongside improvements to the agentflow UI and MCP support.

Apr 14, 2026

Adds support for llama-cpp-server backend and AMD/Intel GPU hardware for PaddleOCR-VL, while fixing dependency issues for NPU/XPU and removing page limits.

Apr 14, 2026

Bug fix release for the platform, resolving issues with applications customization and controlled document printing.

Apr 14, 2026

A major feature release adding 2FA, guest permissions, card layout modes, automation flags, and various editor/desktop improvements.

Apr 12, 2026

This release introduces AST-aware code chunking using tree-sitter, new benchmarking tools, and per-collection model configuration. It also includes numerous fixes for embedding stability and cross-platform compatibility.

Apr 5, 2026

Adds audio-to-video generation capabilities and array batch mode to the workflow tool. Includes significant fixes for undo/redo memory usage and batch execution cancellability.

Apr 1, 2026

Introduces a dedicated Video Input node, full undo/redo history, and Veo model parameter controls. Fixes hydration issues for selected-node execution.

Mar 30, 2026

Adds a Workflow Browser for managing saved workflows, externalizes media files, introduces optional inputs with skip propagation, and redesigns the Group Context Menu.

Mar 29, 2026

Focuses on agentflow improvements, including Rich Text editing, variable system support, and various model and tool fixes.

Mar 23, 2026

Optimizes build times and improves documentation deployment automation, alongside fixing a version mismatch in package files.

Mar 21, 2026