Mar 18, 2025
Last updated
Last updated
Users can now select and delete multiple nodes on the canvas simultaneously, eliminating the need to delete them one by one.
Multi-Selection: Users can select multiple nodes using Ctrl+Click or Right-click (Windows) or Cmd+Click (Mac).
Bulk Delete Option: A Bulk Delete option appears once nodes are selected, allowing users to remove all selected nodes at once.
This enhancement significantly improves efficiency in managing nodes on the canvas.
We’ve migrated our platform documentation to Gitbook to provide users with a more seamless experience and an improved look and feel. The new platform offers enhanced search functionality, making it easier than ever to find the information you need.
Additionally, the tool is powered by AI, allowing users to ask questions directly through the search box and receive quick, relevant answers from our bot.
You can now access the updated documentation at . We hope this transition enhances your experience and makes finding the right information faster and more intuitive.
Users can now duplicate recipes directly on the canvas via a right-click option. The Duplicate option allows users to easily copy the selected recipe and use it within the canvas.
Behaviour Based on Recipe Type:
AskAI Recipe: Requires a chat refresh to re-run previous queries and add the recipe code to the pipeline.
Rapid Model Recipe: Needs to be manually executed after duplication.
Code Recipe: The entire Code tab and recipe structure are copied.
Template Recipe: Displays the same transformation, but users must update the output dataset name before execution.
This feature simplifies the reuse of existing recipes and enhances workflow efficiency.
The platform now displays the data types of each column for intermediate datasets generated after running a recipe.
Key Enhancements:
Schema Button: Now available for datasets generated after running a prompt in AskAI.
Detailed View: Provides column names and their corresponding data types.
This feature improves visibility and understanding of intermediate datasets, making data analysis more efficient.
To ensure environment consistency, the Prediction Service environment will automatically be applied to all queries in the corresponding DataApp in AskAI, eliminating manual configuration.
Users can now use a broader range of naming conventions for Prediction Services.
Naming Rules:
Must start with an alphabet.
Can include numbers (0-9) and underscores (_) after the first character.
This update increases flexibility in naming Prediction Services.
The table component has been standardized across multiple platform pages, including:
Connectors
Dashboards
DataApps
Environments
Project List View
Prediction Service History
Models Page
Additionally, users can now adjust column widths for better usability.
The DataApp creation flow has been enhanced for improved navigation.
Dropdown for Input Type Selection instead of radio buttons.
Cancel Option during creation to stop and return to the DataApps list.
Updated Edit DataApp Page:
Details Tab: Update DataApp details.
Settings Tab: Manage model selection, team access, and response display.
Starter Prompts Tab: Configure starter prompts.
Quick Navigation: Direct access to DataApp from the Edit page.
Default Settings:
Default model set to OpenAI GPT-4o.
Team access enabled by default.
Show model response code toggle enabled by default.
These improvements simplify DataApp management and improve clarity.
To enhance clarity and traceability, the system now generates meaningful names when saving dataset outputs in AskAI or DataApp chat threads.
Several updates have been made to streamline data handling:
Recent Datasets First: The input selection list prioritizes the most recently generated datasets.
Multiple Table Selection: Users can now select up to four tables at a time.
Auto-Deselect Feature: If a user attempts to pin a fifth table, the first pinned table will be automatically deselected with a notification.
These changes simplify dataset selection and prevent exceeding limits.
The sample syntax in Code Recipe outputs has been streamlined to reduce complexity.
Previously required 4-5 lines of code, now reduced to one line.
Applies to all output types.
This improvement enhances readability and efficiency.
The Source and Segment tabs have been restructured for better navigation.
Tabs have been moved to a drop-down menu instead of being displayed separately.
Segment View: Displays segment details when selected.
Source View: Displays dataset source details when selected.
These updates reduce UI clutter and improve accessibility.