v0.1.28

🚀 Release Highlights — v0.1.28 (2025-12-22)

🖼️ Unified Multimodal Tools

• Consolidated Tools: read_media for understanding and generate_media for generation (images, audio, video)
• Backend Selector: Automatically routes to appropriate provider (Gemini, OpenAI, OpenRouter) based on availability
• Multi-Provider Support: OpenAI (gpt-image-1, Sora, TTS), Google (Imagen, Veo), and OpenRouter

📄 Web UI Artifact Previewer

• Document Preview: Preview PDFs, DOCX, PPTX, XLSX directly in the web interface
• Media Preview: View images, HTML, SVG, and Mermaid diagrams inline
• Code Preview: Sandpack-powered code preview with syntax highlighting
• New Components: ArtifactPreviewModal and InlineArtifactPreview for flexible rendering

🔍 OpenRouter Model Filtering

• Tool-Capable Models Only: Model list now filters to only show models supporting tool calling
• Capability Detection: Checks supported_parameters for "tools" capability before including models

---

📖 Getting Started

• Quick Start Guide: Try the new features today
• Try These Examples:
  • unified_multimodal.yaml - Try the unified multimodal tools for understanding and generation
  • massgen --web - Try the Web UI with artifact previewer

What's Changed

• feat: Restricting openrouter models with no tool call support (MAS 165) by @shubham2345 in #680
• fix: Fix Azure OpenAI tool calls and workflow integration by @AbhimanyuAryan in #675
• feat: Standardize multimedia understanding and generation, preview of media in webui, background shell docker fix by @ncrispino in #678
• docs: docs for v0.1.28 by @Henry-811 in #682
• feat: v0.1.28 by @Henry-811 in #681

New Contributors

• @shubham2345 made their first contribution in #680
• @AbhimanyuAryan made their first contribution in #675

Full Changelog: v0.1.27...v0.1.28