This guide explains how to incrementally add text-based information in bulk to your AI Employee's memory. You will learn how to append new data, verify the update history, and reset the text memory when necessary.
RAG text append overview
The RAG (Retrieval Augmented Generation) Text Append feature allows you to add knowledge to your AI Employee without overwriting existing data. Unlike standard updates that might replace a file, this process works incrementally.
When you add text, the system asynchronously converts it into structured topics and tags them with rag_autogenerated_from_text in the AKB. This ensures your AI Employee can access both new and old information based on relevance to the conversation.
Add text to memory
To add new information, you must locate the specific text append attribute in the Builder or Portal.
Navigate to the Builder > Attributes (or the Portal > Settings > Knowledge Base tab).
Locate the attribute
project_attributes_rag_knowledge_base_append_text.Paste the text you want the AI Employee to learn into the value field.
Save your changes.
Once saved, the system processes the text asynchronously. This usually takes a few minutes, though larger volumes of data may take longer.
❗❗ IMPORTANT
The input limit is approximately 100,000 characters. Exceeding this limit may cause processing errors.
Upon successful processing, the system automatically clears the field to indicate readiness for new input.
Verify update history
You can verify that your text was successfully processed by checking the log attribute. This serves as an audit trail and a backup of added information.
Locate the attribute
project_attributes_rag_knowledge_base_append_text_log.Review the content. Newest entries appear at the bottom.
The log uses a specific format to separate entries:
Separator:
---Header:
📅 Append Log Entry [timestamp]
🗒️ NOTE
This log has a maximum capacity of ~180,000 characters. When the limit is reached, the oldest entries are automatically removed.
Reset text memory
If you need to clear all text-based knowledge (for example, to remove outdated pricing or procedures) you can perform a reset.
Locate the attribute
project_attributes_rag_knowledge_base_append_text_reset.Set the value to
True.
This action immediately deletes all knowledge topics tagged as rag_autogenerated_from_text. The attribute will automatically revert to False once the reset is complete.
🚨 WARNING
This action cannot be undone for the active memory. However, your history remains preserved in the project_attributes_rag_knowledge_base_append_text_log attribute if you need to recover data.
Reset and update workflow
If you provide new text in the append field and set the reset switch to True simultaneously, the system will:
Clear the old memory first.
Process the new text immediately after.
Example scenario: Coffee shop menu update
Context: A coffee shop manager needs to add a seasonal "Winter Spice Latte" to the AI Employee's knowledge without removing the standard menu.
Action:
The manager copies the new drink details (ingredients, price, allergens).
They paste the text into
project_attributes_rag_knowledge_base_append_text.After 5 minutes, they check
project_attributes_rag_knowledge_base_append_text_logand see the new entry with today's timestamp.The AI Employee can now answer questions about the "Winter Spice Latte" while still remembering the standard "Cappuccino".