Date: 30th April 2025
We’re excited to roll out Version 1.2.2 of Kadal AI Workbench — focused on delivering deeper AI control, smarter interactions, and enhanced performance for file handling and reliability.
New Features
Interactive Feedback on Agents
Users can now like or dislike AI responses and share specific feedback on accuracy, tone, and relevance — directly improving how Agents evolve over time.
New AI Models Now Supported
OpenAI’s o1 and o3 mini models are now available for Threads and Agents. These can be enabled on request to better align with specific use cases.
Improved File Handling and Stability
File ingestion is now faster and more reliable, especially for large documents. Enhancements include improved chunking, real-time job tracking, and automated retries for fewer failures.
Coming Soon
We're actively working on even more exciting features, including:
- Integrated AI-powered Editor
- Quick agent creation from prompts
- Web search integration for agents
- Support for more LLM model
- Kadal Framework for unified experience