Getting started
Upload source materials for your chatbot to be trained on. Choose from urls, documents, or plain text. If your chatbot doesn't answer a specific question you can add more source data here to handle that here.
General:
When uploading source materials for chatbot training, there is a character limit. Optimal training results are usually achieved with well-structured source data. Ideally, the data should be organized in formats like question/answer pairs or grouped by specific topics.
Web Source Upload:
By providing a base URL, the system will automatically perform a recursive search for all web pages linked to that base URL. The text content from these pages is then collected to enhance your chatbot's knowledge base. For extensive websites, this process may require several iterations, with each search handling up to 50 pages.
Please note that pages rendered using dynamic JavaScript won't be captured in this process. To include content from such pages, use the Text Source Upload method.
Example 1:
Targeting all URLs under https://www.parallelchat.com/
Example 2:
Targeting all URLs under https://www.parallelchat.com/pricing
as the base URL.
Doc Source Upload:
You have the option to upload various document files for chatbot training. The currently supported formats include .doc, .docx, .pdf, and .txt.
Text Source Upload:
This feature allows for the uploading of plain text content. For example, any specific questions your customers have that need answering that might not be available from the other sources. Question/answer pairs works best as well as focusing on a specific topic per text upload.