clai

README

clai: command line artificial intelligence

clai integrates AI models of multiple vendors via with the terminal. You can generate images, text, summarize content and chat while using native terminal functionality, such as pipes and termination signals.

The multi-vendor aspect enables easy comparisons between different models, also removes the need for multiple subscriptions: most APIs are usage-based (some with expiration time).

Features

Piping into LLM:

Easily configurable profiles (note the built in tools!):

Conversation history and simple TUI to browse and continue old chats:

These three features combined makes it quite portable. Personally I have customized my neovim setup to pipe/glob buffers into it, alongside pre-prompted profiles, to create what I call GhettoPilot. But that's besides the point. Let's continue:

Supported vendors

• OpenAI API Key: Set the OPENAI_API_KEY env var to your OpenAI API key. Text models, photo models.
• Anthropic API Key: Set the ANTHROPIC_API_KEY env var to your Anthropic API key. Text models.
• Mistral API Key: Set the MISTRAL_API_KEY env var to your Mistral API key. Text models
• Deepseek: Set the DEEPSEEK_API_KEY env var to your Deepseek API key. Text models
• Novita AI: Set the NOVITA_API_KEY env var to your Novita API key. Target the model using novita prefix, like this: novita:<target>, where <target> is one of the text models.
• Ollama: Start your ollama server (defaults to localhost:11434). Target using model format ollama:<target>, where <target> is optional (defaults to llama3). Reconfigure url with clai setup -> 1 -> <ollama-model-conf>

Note that you can only use the models that you have bought an API key for.

Get started

Install Glow for formatted markdown output when querying text responses.

go install github.com/baalimago/clai@latest

You may also use the setup script:

curl -fsSL https://raw.githubusercontent.com/baalimago/clai/main/setup.sh | sh

Either look at clai help or the examples for how to use clai.

Honorable mentions

This project was originally inspired by: https://github.com/Licheam/zsh-ask, many thanks to Licheam for the inspiration.