Using GPT Pilot with Anthropic models

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 the How to start using GPT Pilot article to setup GPT Pilot via the terminal.

1 - To use Anthropic, get an API key from the Anthropic website.

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. If you're using Anthropic via a proxy or through AWS Bedrock, you'll also need to set base_url accordingly from the default null:

"llm": {
  "anthropic": {
    "base_url": "https://api.anthropic.com/",
    "api_key": "your-anthropic-key",
    "connect_timeout": 60.0,
    "read_timeout": 10.0
  }
},

4 - Update the agent section of the config.json file for Anthropic. Note: This is an example, you'll need to double-check the official Anthropic documentation for the most recent model info:

"agent": {
    "default": {
      "provider": "anthropic",
      "model": "claude-3-5-sonnet-20240620",
      "temperature": 0.5
    }

5 - After making changes to your config.json file, if you're using the VS Code extension, you will need to restart VS Code for the changes to be reflected in the settings page.

6 - That's it!