llm-sofa-company

Description

This is a simple app to display a list of sofas and their details. The app hooks into an LLM to take a description and output a list of tags to search for. This currently can use either OpenAI or AWS Bedrock - Claude.
See Environment Variables section on how to configure.

Installation

Install dependencies

npm install

Environment Variables

This uses direnv to manage environment variables. You will need to create a .envrc file in the root of the project See .envrc.example for an example of the required variables

Run the app

npm start

Usage

Navigate to http://localhost:3000 to view the app.

Update the prompt to test around diffent prompts. See llm.setTemplate in server.ts.

Observability

This app uses OpenTelemetry to send traces to Honeycomb.io by default.