ChatGPT Guide

Axiom can interact with ChatGPT to send any prompt, as well as extract data.

You'll need to add your OpenAI key for each ChatGPT step.

Axiom currently gives the option to use GPT-3.5-turbo and GPT-4 in automations.

# Set limits on your API key

Currently, OpenAI does not allow per user or per API key limits.

One workaround is to create a dedicated account, with API key for axiom, to isolate permissions.

We recommend setting limits on total API spending, rotating API keys periodically, as well as following OpenAI's general security advice:

Read more (opens new window)

# Generate text with ChatGPT

This step will send any custom prompt to ChatGPT, and axiom will read the response into one of its tokens.

It's possible to use axiom tokens within the prompt sent to ChatGPT to make dynamic prompts.

axiom.ai and Chatgpt generating text in the browser

# Extract data with ChatGPT

This step allows you to extract data from text with ChatGPT.

Some examples provided include extracting product data, or profile information from social media.

Other common examples include, e-mail addresses, or postal addresses, but you can use this step to instruct ChatGPT for any type of data you wish.

axiom.ai and Chatgpt scraping and extracting data

# Use conditional logic with ChatGPT

It's possible to use conditional logic with ChatGPT, by passing data from a 'Generate text with ChatGPT' step into a 'Conditionally jump' step.

This may be useful if you'd like ChatGPT to make a decision, then perform certain actions.