Data
Last updated
Last updated
We support the following data sources:
You are able to select any text data you uploaded to the Data Source in your workspace. Vector database can help you retrieve the most important segments of a long text for an LLM. For example, you can upload a text document to the vector database. When you search some text segments in the vector database, it will return the most relevant text segments.
Limit: Return the top n search results.
Query: Input (text segment) you want to search in the vector database. Can be either dynamic (entered user input) or fixed value (set to a constant value).
This component can help you create any text, image, or document inputs. For example, a text Resource Loader is useful to create an instruction.
Resource loader can be either dynamic (entered user input) or fixed value (set to a constant value).
Chat history component allows you to use any chat history data source you saved for the selected projects. This is useful to search for anything you talked with the selected bot.
Limit: Return the top n search results.
Query: Input (text segment) you want to search in the vector database. Can be either dynamic (entered user input) or fixed value (set to a constant value).
This component will search the Google for a given search query and return the first result from the top 3 searches. This block is really useful for getting up-to-date / live information that some AI models might not have access too.
YouTube Database component allows you to use the YouTube video content uploaded in the Data Source. The subtitles of the video was extracted and save to the vector database. Therefore, you can use this component to quickly search for any segments in the videos.
Must upload videos that have subtitles.
Limit: Return the top n search results.
Query: Input (text segment) you want to search in the vector database. Can be either dynamic (entered user input) or fixed value (set to a constant value).
