Build an LLM Agent in 5 minutes

Overview

This repository shows you how to build an LLM using Snowflake's API.

Instructions

1. Clone this repo.

2. Sign up for an account here. Activate your account and you should see a welcome screen.

3. Go to Projects -> Worksheets -> Click Create SQL Worksheet

4. Copy and paste the entire setup.sql into the worksheet and hit run. It might take a couple of minutes for Step 9 to finish! You should see the following response: Stage area MODELS successfully created.

5. Go to Data -> Add Data -> Click Load files into a Stage.

6. Upload the sales_metrics_model.yaml file. Make sure to select SALES_INTELLIGENCE.DATA as your database + schema and MODELS as your 'Stage'.

7. Create a folder called .streamlit and a file called secrets.toml inside.

8. Click on your name in the bottom left corner, and select Connect a tool to Snowflake. Use the dialog to fill out your information, replacing values in [] like so:

[snowflake]
account = "[Account Identifier]"
user = "[User Name]"
password = "[Same as when you signed up. Be careful here and don't check this in to GitHub!"
role = "[Role]"
host = "[Account/Server URL]"

9. Run pip install -r requirements.txt to make sure you have all the dependencies working.

10. Run streamlit run streamlit.app and you should see a chat assistant ready to work with you! Try a few prompts:

• "What was the total sales volume last year?" should output a SQL query and an interpretation of your request.
• "Summarize the call with TechCorp Inc" should give you a summary of the call transcript.