February 2025

Core Features

Added powerful Node.js execution capabilities to the application, allowing users to run custom JavaScript code directly within the platform
Integrated browser-based speech-to-text capability to the chat screen, allowing users to dictate messages instead of typing
Implemented comprehensive licensing structure for better subscription and access management
Added in-memory queue for messaging that supports 500 concurrent questions and 1000 questions per minute, preventing system overload during high traffic

Enhanced App steps with Clone button, collapsible structure, and Drag and Drop functionality
Added robust data persistence for steps during application updates
Added powerful new steps to the App section:
- SQL Executer: Run SQL queries directly from workflows
- Send Email: Automate email communications
- Bing Search: Retrieve information from Bing search engine
- Bing News Search: Access current news articles
- DuckDuckGo Search: Alternative search engine integration
- Scrape Web Page: Extract content from websites
Added file upload capability for Apps to process document-based workflows
Added JSON output option for LLM steps, enabling structured data extraction from language models

Fixed mobile warning designs for better readability and user experience
Added Dashboard to mobile TabBar for easier access to key metrics
Streamlined mobile interface by removing redundant menu items from SideBar
Enhanced App section's responsive design for better usability on smaller screens
Fixed navigation menu issues in the App section when accessed from mobile devices

Added Nova Credit display to the chat screen for real-time credit monitoring
Redesigned Nova Credit card with improved visual aesthetics and clearer information display
Added more user-friendly alerts for Nova-related notifications and status updates
Activated $200 Nova Credit addition feature for users to expand their usage capacity

Updated GDPR section in the database (previously labeled as KVKK) for better regulatory alignment
Improved GDPR compliance by requiring checkbox selection before signup button activation

Prepared Docker deployment for production environments, enabling easier installation and management
Added ability to configure Embedding model in settings for customized vector embeddings
Added option to enable/disable multi-tenant functionality based on deployment needs
Implemented automatic startup with admin credentials from environment variables
Added Azure OpenAI services integration
Added default model selection for application-wide use

Enhanced message formatting for copy-pasted content, preserving original formatting and improving readability
Updated badge color to blue for synchronized datasources for better visual indication of sync status
Added intelligent space cleaning function for API Key and other copy-paste fields to prevent errors from trailing spaces
Added default false setting for checkbox types in App step forms for more predictable behavior
Added search functionality to the Peer list page for quicker access to specific Peers
Fixed various bugs throughout the application for improved reliability
Fixed state issues in the Workspace Embedding Model settings for more reliable configuration
Resolved slow performance issues in the App editor, significantly improving responsiveness
Fixed state issues in App editor for more reliable editing experience
Resolved error when adding steps with duplicate output names in App editor

Fixed general state and design issues in the Schedule section
Added configurable limits to scheduler for datasources and apps (default: 10 total, adjustable via MAX_JOB_LIMIT environment variable)