ChatGPT - Generate Response
Overview
Sends a prompt to ChatGPT and returns the message. The results will be printed in the console and saved in a text file. If you generate multiple responses, the file name will have an added suffix.
Remember: ChatGPT should be used as a starting point, not a replacement for a human in the process. Use it to rough draft content, summarize insights, or jumpstart repetitive work—but always review and tailor output before use.
Best Practices is don’t send ChatGPT-generated content directly to clients without human review.
For example, instead of sending GPT generated insights directly to the client:
→ Try slacking the file to your team channel for feedback
→ Or uploading to Google Drive and review together before sharing externally
Always ensure what’s shared reflects our standards, expertise, and strategic thinking.
Want to pass a CSV, TXT or MD file to ChatGPT in your prompt? See How-To: Embed File into a Template Input for instructions on how to do so,
Variables
Name | Reference | Type | Required | Default | Options | Description |
|---|---|---|---|---|---|---|
Prompt | CHATGPT_PROMPT | Alphanumeric | ✅ | - | - | Prompt that will be sent to ChatGPT |
Local File Name | CHATGPT_DESTINATION_FILE_NAME | Alphanumeric | ✅ | - | - | The name for the txt file once it is generated. If more than one response is generated, the files will have counted suffixes (file.txt, file_1.txt,...) |
YAML
Below is the YAML template for this Template and can be used in the Fleet YAML Editor.
source:
blueprint: ChatGPT - Generate Response
inputs:
CHATGPT_API_KEY: null ## REQUIRED
CHATGPT_RESPONSES: '1' ## REQUIRED
CHATGPT_RANDOMNESS: '1'
CHATGPT_PROMPT: null ## REQUIRED
CHATGPT_DESTINATION_FILE_NAME: null ## REQUIRED
type: BLUEPRINT
guardrails:
retry_count: 1
retry_wait: 0h0m0s
runtime_cutoff: 1h0m0s
exclude_exit_code_ranges:
- '0'