AI Retrieval allows your ai agent to fetch relevant information from your knowledge base based on your query. It allows precise filtering and response formatting for optimal retrieval.
1. Search Query:
Enter a query to extract relevant content from the knowledge base.
2. Tags:
Tags are created in the training section, allowing you to assign specific tags to specific training data. Tags help filter the retrieval process by segmenting the knowledge base.
It will retrieve all your training data that matches the query. These can be used to refine search results:
- Include: Retrieves content that contains at least one of the selected tags.
- Exclude: Filters out documents containing the specified tags.
3. Nodes Limit:
Sets the number of knowledge nodes retrieved per query. A higher limit returns more results, whereas a lower limit narrows the response.
4. Response Type:
Select the type of response format you want to retrieve:
- Raw: Fetches the plain text result.
- JSON: Converts the raw data into JSON based on the provided JSON schema.
5. Store Into:
Allows saving retrieved responses into a variable for further processing or automation.
For more details check out our Help Center article on How to use Retrieval in Studio.