CustomGpt

.NET Core sample application to chat with custom data using OpenAI APIs (Embedding and ChatCompletion)

Database used:

• SingleStore as a vector DB to store embeddings

Application Building Blocks

• Store your data in a vector database (this sample application uses SingleStore)
• Convert user query to vector embedding (using OpenAI API to generate embedding) and perform a similarity search against your pre-processed embeddings
• Feed the search result to LLM (OpenAI Gpt in this sample)
• Display the result

TODO

• Adding support for Supabase vector and Postgres DB with pgVector
• Maybe use a different embedding logic than OpenAI API
• Code refactor