v1.91.2

• 🚀 Enhanced Data Handling: Introduced a unified dataTryParse function for seamless parsing of various data formats, improving flexibility and reducing redundancy.
• 🛠️ New Tools: Added fs_infer_schema and fs_data_query tools to infer JSON schemas and query data using GROQ syntax, empowering data analysis capabilities.
• 📂 Data Agent: Launched a new "data" agent to analyze and query file data, equipped with schema inference and querying tools.
• 🔒 Safety Improvements: Integrated content safety checks for harmful and protected material in agents and tools.
• 🎥 Video Agent Update: Enhanced the video agent with additional safety checks.
• 📄 Teams Integration: Added functionality to read messages from Microsoft Teams channels, complementing the existing message posting feature.
• 🧪 Sample Scripts: Introduced new sample scripts demonstrating data schema inference and multi-file data analysis.

v1.91.1

✨ Enhanced Video Capabilities
Introducing powerful video processing tools! Transcribe audio/video files with speech-to-text, extract frames, probe metadata, and even extract audio from video files. Perfect for multimedia projects!

🌐 WebSocket Improvements
WebSocket server now supports dispatching progress updates to all clients, enhancing real-time collaboration and monitoring.

🎤 Speech-to-Audio Support
Added speech synthesis capabilities, enabling text-to-speech generation with customizable voices for a variety of use cases.

🛠️ Developer Utilities
New deleteEmptyValues utility improves object handling, and CLI enhancements include options for dispatching progress events.

🚀 Performance and Stability
Refactored internal structures and fixed minor issues, ensuring a smoother and more reliable experience.

v1.91.0

🚀 GenAIScript 1.91.0 Highlights

🌟 New Features & Improvements

• Enhanced Embedding Configurations: Added new embedding aliases for streamlined LLM setups.
• Streaming Output: Introduced robust real-time output streaming for better user feedback.
• Trace-Driven Output: Improved trace capabilities, enabling detailed output tracking and customization.

🛠️ Quality Enhancements

• Model Management: Simplified model pulling with expanded support for LMStudio and enhanced alias handling.
• Video Config & Fetch Updates: Improved video configuration documentation and fixed fetch command formatting for better usability.
• Error Handling: Enhanced error serialization for clearer debugging.

🔧 Developer Experience

• Refined APIs: Modularized chat message collapsing and moved it to a dedicated rendering module.
• Expanded Trace API: Added new output trace capabilities, including detailed logging and structured output support.

🧹 Maintenance

• Code Cleanup: Removed unused code paths, updated aliases, and optimized internal logic for better performance and readability.

v1.90.1

🎉 Enhanced Video Utilities: Added support for WorkspaceFile in video processing functions, enabling seamless handling of various file types. Extracting audio, frames, and metadata from videos is now more robust and flexible.

✨ Improved Fetch Implementation: Integrated cross-fetch for better compatibility and enhanced proxy support. Added detailed request tracing, including support for multipart form data, to simplify debugging.

🔧 Smarter Configuration Handling: Configuration merging is now more intuitive, with support for arrays and object properties like modelAliases. This update improves flexibility and reduces potential errors.

🐛 Bug Fixes: Resolved issues with error handling, including a fix for misreported error messages. Enhanced the reliability of OpenAI API interactions and transcription workflows.

📂 Sample Configurations: Introduced sample genaiscript.config.yml and large.genai.mjs files to demonstrate configuration and scripting capabilities.

v1.90.0

✨ New Features and Enhancements

• Added CLI commands for video processing:
  • Extract audio from video files.
  • Extract video frames with options for count, size, and output folder.
• Integrated HuggingFace's pipeline API for advanced transformer capabilities.
• Introduced transcription support via OpenAI's Whisper API, enabling audio-to-text conversion with caching and SRT/VTT format generation.
• Enhanced data slicing in defData to support object field sampling and filtering.

⚡ Performance Improvements

• Optimized hashing with streaming file support and salt integration.
• Improved concurrency handling for video frame extraction and audio transcoding using FFmpeg.

🛠️ Bug Fixes and Stability

• Fixed edge cases in file handling for workspace paths.
• Resolved hashing inconsistencies for various data types like buffers and blobs.
• Enhanced error handling and logging for video processing and transcription workflows.

🎥 Video and Audio Processing

• Seamless integration of FFmpeg for video/audio tasks.
• Added caching for video probes, audio extraction, and frame generation.

💡 Developer Experience

• Simplified runtime configuration for transcription and video utilities.
• Improved CLI usability with detailed descriptions and argument validations.

v1.89.2

🚀 What's New in GenAIScript 1.89.2

• Enhanced Hash Functionality: The hash function now supports additional input types like ArrayBuffer, Blob, and null/undefined values. Improved flexibility ensures seamless handling of diverse data structures. Comprehensive tests included for reliability. 🔒✨
• Filtered Script Listings: Unlisted scripts are now excluded from project listings, offering a cleaner and more focused view of available scripts. 🎨
• Improved Error Logging: Added error logging during model pulls to help identify and debug issues more effectively. 🛠️

v1.89.1

🚀 Enhanced Usability
The RunButton now dynamically updates to display "Abort" when a process is running, making interactions more intuitive.

🤖 Model Updates
Introduced support for the new ollama:phi4 model, offering advanced capabilities. Additionally, updated model aliases for consistency and clarity.

♻️ Streamlined Integrations
Removed the unused DeepSeek base URL for a cleaner configuration. Improved key consistency across integrations, including topP.

v1.89.0

🎉 New Features

• File Selection for Imports: You can now select specific files to include when importing, streamlining your workflow.
• Video Recording in Browser: Added support for recording browser sessions with Playwright, including video file retrieval for analysis or debugging.
• ReadableStream for Images: Enhanced image handling with support for ReadableStream inputs, expanding compatibility.
• Playground Setup: A new playground environment is now available, with live file watching for faster iteration.

🛠 Improvements

• Enhanced Chat Participant Logic: Introduced support for modifying chat messages dynamically during processing.
• Updated Memory Aliases: Memory alias updated from large to small for optimized configurations.
• Markdown Rendering Fixes: Improved handling of complex content arrays in rendered markdown, ensuring better formatting.

♻️ Refactors

• Simplified image processing functions by removing redundant awaits.
• Reordered StatsTabPanel for better UI flow.

📚 Documentation

• Added instructions for running the playground in CONTRIBUTING.md.

This release brings powerful new tools and refinements to enhance your GenAIScript experience! 🚀

v1.88.5

🚀 New Features

• Introduced a powerful Markdown diff utility for comparing text changes, now integrated into the UI for enhanced clarity.
• Added a file existence check for web.mjs.map, ensuring more robust error handling.

🎨 UI Enhancements

• Improved Markdown rendering with better annotation processing and support for GitHub-style alerts.
• Extended schema support to include svg and path tags, enabling richer visual content in rendered Markdown.

🛠️ Developer Experience

• Optimized annotation handling by refining conversion logic and simplifying dependencies.
• Added a new markdownDiff function for generating diff views in Markdown format.

This release focuses on improving usability, reliability, and the overall developer experience. Enjoy the updates! ✨

v1.88.4

✨ Expanded File Format Support
This release enhances file handling capabilities by introducing support for additional file formats, including .txt, .pdf, .md, .ts, and .prompty. You can now seamlessly work with a broader range of files in your workflows.

📂 Improved Workspace File Integration
Workspace files are now better integrated into the file expansion process, ensuring smoother handling and accurate content resolution across your projects.