Working with the Knowledge Base

In order to optimize the performance of your AI chatbot and configure the most accurate and necessary responses for users, you must follow our recommendations when adding documentation. Currently, files can be downloaded in txt, pdf, docx formats. File size limit 1MB. There are no restrictions on the number of uploaded articles.

Best Practices for Adding Documentation

1. In the description of the article, use keywords and tags related to the article. The chatbot will compare the client’s question with article headlines and select the most appropriate one. For example, when adding an article to the knowledge base describing the cost of Cavuer services, you should indicate in the title: cost, language model price, payment, tariff;
2. Add tags and keywords in different languages ​​to the title of the article. The allowed header length is 255 characters;
3. Don't load long text, break large texts into small ones that answer specific queries. So, when processing a question, an article from the documentation will be processed to find the optimal answer to the client’s question. Processing an article to find an answer will be considered as incoming tokens and charged according to the specified prices. For example, information on creating chatbots should be divided into three articles: creating a Widget chatbot, creating a Telegram chatbot, creating a WhatsApp chatbot;
4. When downloading documentation, be sure to set headers carefully. The title cannot be changed in the future. It is possible to change the content of the article.
5. If you have complete documentation on your site, you can upload basic information into articles and attach a link to the documentation. This way, the client will receive the answer he needs, and for detailed information he will be able to follow the link to your documentation;
6. Do not use special characters or emoticons.