Emma
Summary:
-
ChatterDocs stores user data using vectorization in Supabase with row-level security.
-
The chatbot prevents prompt hacking and uses OpenAI's Chat-GPT API for natural language processing.
Key terms:
-
Vectorization: A process that allows AI to search through content for relevant answers based on the questions asked.
-
Supabase: A platform where data is stored with row-level security.
-
Prompt hacking: Exploiting a chatbot's response generation process to obtain undesired or inappropriate answers.
-
Chat-GPT API: OpenAI's natural language processing model used by ChatterDocs for understanding and responding to user queries.
-
Row-level security: A security measure that restricts access to data based on user permissions.
Tags:
ChatGPT
Research
OpenAI
Tools
GPT-3
AI chatbot
AI models
Chatbot
Integration
Notion