- Smartly.AI :
- Virtual Agent :
We have introduced a new feature that allows you to easily manage your API keys. With this update, you can create, edit, or delete API keys based on your specific requirements. This enhances flexibility and security by enabling the generation of distinct keys for different environments or implementing temporary keys.
We have added new endpoints to the API that allow you to manipulate your Node.js code. The available endpoints are as follows:
- Get all custom codes: API Reference
- Get one custom code: API Reference
- Update a custom code: API Reference
This API provides convenient ways to interact with your custom code directly from the API.
We have introduced a new API that enables you to manipulate rich messages. The available endpoints are as follows:
- Get all rich messages: API Reference
- Get one rich message: API Reference
- Update a rich message: API Reference
- Delete a rich message: API Reference
- Create a rich message: API Reference
This API allows you to easily manage and customize the rich messages in your application.
New: Generative AI duplication
You can now duplicate Generative AIs, allowing for easier experimentation and iteration.
New: Improved list view
In the list view, we now display the training status for each generative AI, so you know when an action is needed from your side. We have also added the creation and last change dates for better visibility.
New: Revise Gen AI answers
If you come across an incorrect response from the Generative AI while reviewing conversations, you can now easily edit the response. This ensures that the AI learns from its mistakes and provides accurate responses in the future.
New: OCR on scanned documents
We have introduced Optical Character Recognition (OCR) for scanned documents. OCR allows us to extract relevant content from documents that are in the form of scans or photographs. Supported formats for OCR include pdf, jpg, jpeg, and png. The OCR feature is automatically utilized when necessary, requiring no manual action.
Improved: Smart Splitting
We have improved the smart splitting mechanism for documents that require more attention to be chunked properly. By utilizing a semantic approach to chunking, the structure and meaning of the document are preserved, resulting in more accurate results.
Improved: Data source management
We have introduced the ability to scan web pages and extract all the links as new data sources. Additionally, we have simplified the management of data sources. When uploading files, the document titles are now displayed for easier identification.
The generative AI mode is now available for all users. However, to access this mode, you need to complete a form. Additionally, some advanced settings are hidden behind the "Advanced options" button for a cleaner user interface.
Ping us to get access to this feature.
Improved: Votes on bot answers are now displayed on the Conversations pages. Supported votes include those coming from the webchat and the API, providing valuable feedback to improve the bot's responses.
Fixed: A webhook issue with the Messenger integration has been resolved, ensuring smooth communication between the bot and the Messenger platform.
Fixed: An issue with expired image links has been addressed, allowing for proper rendering of images in the bot responses.
Fixed: Fixed bug causing discrepancies in displayed time between the platform and exported log files, ensuring synchronization of time in the Conversations tab and Excel export.