terminal-ai

This is a simple CLI tool for chatting to an LLM in the terminal.

Currently, only Gemini is supported, but I'll add Ollama support later. Nothing else is planned for now.

---

Command overview

Consider this a simple overview of some of the commands. The commands have more functions, which can be viewed by typing chat -h or just typing chat. There are also some additional commands not mentioned here

• chat start - Begin a chat
• chat once <message>- Send a single message to the LLM
• chat list - List all existing chats
• chat delete <chat name> - Delete a chat
• chat systemprompt <...> - System prompt configuration
• chat model <...> - Model configuration (Add models and API keys here)

When in a chat using chat start, there are some in-chat commands. These are:

• quit/bye/exit to quit (saves your chat)
• clear to clear the chat history (asks for confirmation)
• retry to regenerate a new AI response to the previous message
• save to save the chat so far
• systemprompt/system to display and/or change the system prompt
• help to display a help message

Enjoy :)