Fast custom prompts to GPT-3, GPT-4 and ChatGPT API
An extension to have fast and custom prompts sent to OpenAI's GPT-3, GPT-4 and ChatGPT from every webpage.
Ensure the lowest cost for using GPT querying directly OpenAI`s API.
The extension works using your API Key from OpenAI.
Get one at: https://beta.openai.com/ (we never have access to it).
This is an open source project available at GitHub: https://github.com/giosilvi/GPT-Prompter
***How to use it***
(On first install you need to reload the open webpages to have it working)
1. Click on the extension icon and insert your API KEY.
2. (optional) Select text on a web page.
3. Right-click, from the context menu select GPT-Prompter and choose a prompt!
Complete Guide: https://www.gptprompter.com/how-to-use-it
You can customize the prompt as you like, but they always must contain the keyword #TEXT#. It gets replaced with the text you select from the webpage.
Multiple models with different capabilities and costs are available from OpenAI:
- gpt-3.5-turbo-intruct (substitute davinci, curie,etc.)
***History and costs***
You can access and search through the history of your prompts and save them as JSON file.
v4.0: Added Instruct model, and remove stream for message (to make it work with v1/completions)
v3.4: Whisper API added. Codex model deprecated.
v3.3: Gpt-4 integration. Removed max tokens, now tokens are automatically computed from prompt.