The cyclone symbolizes the idea of a pipe that connects between elements, while the speech balloon symbolizes the idea of the user being able to generate prompts and receive AI-generated responses. Together, they represent the library's ability to empower users to execute code or view diagrams directly from the command line, enhancing productivity and creative workflows.
Prompt-Flow is a library that enables seamless integration of any conversational AI model into command-line interfaces (CLIs) through piping. It empowers users to generate prompts, receive AI-generated responses, and execute code or view diagrams directly from the command line, enhancing productivity and creative workflows.
To install Prompt-Flow, cd into the directory and run npm install. To configure it, create a openai_credentials.json file in the root folder, with two fields: apiKey and orgId from your OpenAi account.
After installing and configuring Prompt-Flow, you can take advantage of 'npm run executables' to run commands (from everywhere) in this way:
echo 'hello world with a great twist! ' | code | ai | execute
instead of
echo 'hello world with a great twist! ' | ./engine/code.js | ./engine/ai.js | ./engine/execute.js
in the directory of installation.
- Run the command
npm run executablesin your terminal. - Ensure that your system's PATH includes the directory where the Prompt-Flow executable resides (/engine).
echo 'hello world with a great twist! ' | code | ai | executeecho 'different types of coffee' | mermaid | ai | view-mermecho 'how to make coffee' | markdown | ai | view-mdecho 'plot these countries and their populations: israel, lebanon, egypt, greece | chart | ai | executeecho 'how to make coffee' | html | ai | openecho "using leaflet js library, show on a worldmap a sample of 10 random countries with their population count; include the leaflet script as part of the html\!" | html | ai | browserecho "how to make coffee mindmap" | markdown | ai | view-mindmap