Using GPT Pilot with OpenAI

Note: This article is for using GPT Pilot locally via the terminal. For info on how to use your own API key via the Pythagora VS Code extension, read this article.

0 - Be sure to first follow the steps in How to start using GPT Pilot from the GPT Pilot README to setup GPT Pilot via the terminal.

1 - Go to the OpenAI website and get an API key.

2 - If you haven't already, in your cloned GPT Pilot repo, make a copy of the example-config.json file and name it config.json.

3 - Edit the new config.json file's llm section and remove the other providers that you're not using:

"llm": {
  "openai": {
    "base_url": "https://api.openai.com/v1/",
    "api_key": "sk-your-api-key",

4 - Update the models section of the config.json file for OpenAI (if not already set as openai by default). Note: This is example code, you'll need to double-check the official OpenAI documentation for the most recent model info:

"agent": {
    "default": {
      "provider": "openai",
      "model": "gpt-4o-2024-05-13",
      "temperature": 0.5
    },