aaronccasanova / markdownchat

Markdown chat conversations with Large Language Models

markdownchat

A tool for facilitating markdown chat conversations with Large Language Models.

Quick start

1. Add an OPENAI_API_KEY to the current environment.

export OPENAI_API_KEY=<api-key>

2. Initialize a markdownchat file in the current directory.

npx markdownchat

3. Add a chat message to the end of the created markdownchat.md file.
4. Re-run the tool to stream responses back to the markdownchat file.

npx markdownchat

That's it! You can continue the markdownchat conversation by repeating steps 3 and 4.

Important: Keep in mind the cost increases the longer the conversation becomes. It's recommended to rename the file to save the conversation and start a new markdownchat!

Usage

Usage
  $ npx markdownchat [markdownchat-file] [options]

Options
  -h, --help         Show help text
  -v, --version      Show version number
  -m, --model        Name of model (default: "gpt-3.5-turbo")
  -t, --temperature  Temperature of model (default: 0.5)
  --env-file         Path to .env file

Examples
  $ npx markdownchat
  $ npx markdownchat ./markdownchat.md
  $ npx markdownchat ./markdownchat.md --model gpt-4.5-turbo

Environment Variables

• OPENAI_API_KEY: API key for OpenAI API
• OPENAI_BASE_URL: Base URL for OpenAI API (default: "https://api.openai.com/v1")

These environment variables can be set in a few ways:

• Exporting in the current shell.

export OPENAI_API_KEY=<api-key>
npx markdownchat

• Inlining before the npx markdownchat command.

OPENAI_API_KEY=<api-key> npx markdownchat

• Using an .env file.

npx markdownchat --env-file .env

Contributing

Contributions are welcome. Please submit a pull request or create an issue to discuss the changes you want to make.