#GPT-CLI: Command Line Interface for OpenAI's GPT

7 messages · Page 1 of 1 (latest)

rustic pagoda
#

The primary goal of this project is to create a command line interface (CLI) for GPT, allowing you to access its capabilities right from your terminal.

Example:

gpt prompt "create a bash Hello World script" > hello.sh
cat file1.txt file2.txt | gpt prompt "combine and summarize the information from these two texts" model "text-davinci-edit-001" > summarized_information.txt

https://github.com/kainazzzo/GPT-CLI

GitHub

Contribute to kainazzzo/GPT-CLI development by creating an account on GitHub.

sacred spindle
#

That seems super useful! Will definitely try it out 🙂

rustic pagoda
rustic pagoda
#

modified it to use chat completions, which simplifies things a lot.. it was using the v1 text completions api, which made it difficult to use, because you had to specify the text-davinci-003 or code-davinci-002 model. Now it just uses the chat format to describe the desired way to combine standard input and prompting.

rustic pagoda
#

I fixed up the command line syntax to be more POSIX friendly.. getting close to 1.0!

rustic pagoda
#

Added embeddings support! Now you can create and save embeddings and use them to add context for prompts & chat

#
scrapeUtil http://yoururl | gpt embed > yourUrl.dat
# then
gpt chat --file yourUrl.dat
# or
gpt --file yourUrl.dat --prompt "Ask a single question with yourUrl.dat as context."