Ask The Documents (Embeddings / RAG / ChatGPT) with Wasp & PG Vector

Live at https://ask-the-documents-client.fly.dev/

What does it do?

This is an example Wasp app that supports:

• scraping whole link hierarchies (great for docs)
• scraping a single link
• generating embeddings for page content
• semantic search using PG Vector
• chatting with the documents using OpenAI's ChatGPT

Running it locally

1. Make sure you have the latest version of Wasp installed.

   bashCopy

   npm i -g @wasp.sh/wasp-cli@latest
   

2. Start the database with:

   bashCopy

   wasp start db --db-image pgvector/pgvector:pg18
   

3. Copy .env.server.example to .env.server and fill out the environment variables.

   bashCopy

   OPENAI_API_KEY="<your_openai_key>"
   GOOGLE_CLIENT_ID="<your_google_client_id>"
   GOOGLE_CLIENT_SECRET="<your_google_client_secret>"
   

4. Migrate the database with:

   bashCopy

   wasp db migrate-dev
   

5. Start the server with:

   bashCopy

   wasp start