Chat Data

What does Chat Data do?

Chat Data is an AI agent platform for customer support chatbots and live chat on websites. You can build an AI agent from your documents or site content, then embed a chat widget or API so the agent answers questions from your knowledge base.

Where is user(my) data physically stored?

Our servers are located in Switzerland, so users' data is physically stored within Swiss borders and hosted on Oracle Cloud infrastructure. Additionally, file blobs are also stored in Switzerland using Azure Blob Storage services.

How should my data be formatted?

At present, you have the option to upload individual or multiple files in formats such as .pdf, .txt, .html, or .docx. Alternatively, you can paste text or provide a website link for data scraping.

Can I integrate your chatbot widget with my custom model backend?

Absolutely. Employ our flexible chatbot widget, effortlessly embeddable anywhere, driven by the backend URL you supply. This method is not only supported but also encouraged, as it minimizes the impact on our computing resources.

Tell me about the Medical Chat models.

The Medical Chat models are pre-trained models developed using Chat Data. It is created with a meticulous selection of data sources curated by the our platform to offer proficient knowledge base specializing in human and veterinary medication. To ensure HIPAA compliance for our medical chat models, we refrain from retaining chat history involving agents within the medical chat models. For more information, please visit https://medical.chat-data.com/

Is it possible to provide instructions to my agents?

Certainly, you have the ability to modify the base prompt, assign a name to your agent, specify its personality traits, and provide instructions regarding how it should respond to questions, such as only responding in French, for example.

What should be included in the base prompt and what should be used as the training data?

The instruction that applies to the entire conversation should be included in the base prompt, while the knowledge point that needs to be retrieved based on the user's latest query should be used as the training data.

What is the process for incorporating my agent into my website?

You can integrate your agent by either embedding an iframe or adding a chat bubble to the lower right corner of your website. To achieve this, create an agent and select the 'Embed on website' option. Alternatively, you can utilize the API to engage with your agent from any location.