June 12
A focused release bringing stronger agent quality, faster execution, greater customization, and improved platform usability.
Agent Builder
Agent Executor 1.3
Released Agent Executor v1.3 with enhancements focused on performance, reliability, and developer experience:
Improved Monitoring & Agent Logs for better execution visibility
Enhanced Evals for pre-deployment quality and reliability testing
Added Context Compaction for improved long-running conversations
Introduced Fast Mode for quicker agent responses
Added OpenAI-Compatible Endpoint support for external application integration
Introduced Document Generation capabilities
Added Skill & Runtime Version Deprecation Framework
Simplified and cleaned up document-related skills
Reinforcement Learning with Human Feedback (RLHF)
Released Reinforcement Learning with Human Feedback (RLHF) capabilities
Improve agent behavior through human feedback and corrections
Approved corrections can be added to agent or organizational memory to improve future responses over time
Additional Improvements
UX improvements across the Agent Playground
Better build and testing experience
Bug fixes and stability enhancements
App Builder
Model & Design Flexibility
Added support for selecting Claude or Gemini model during app creation
Introduced Design Style Selection with options such as Minimal, Neomorphism, Glassmorphism, Modern SaaS, Enterprise and more
Reliability Improvements
Improved user communications during app generation
Better error handling and timeout management
Solutions
Refreshed Experience
Improved Solutions interface with cleaner navigation and richer information for AI Agents and Agentic Apps
Enhanced discovery and evaluation experience for showcased solutions
More improvements are on the way – keep building and sharing your feedback with us.
Happy building! 🚀
Last updated