Yet another Rust library and CLI for interacting with LLMs.
# Basic query rullm "What is the capital of France?" # Use different models with aliases rullm --model gpt4 "Explain quantum computing" rullm --model claude "Write a poem about the ocean" # Use templates for structured queries ({{input}} placeholder is automatically filled) rullm -t code-review "Review this function" # Interactive chat rullm chat --model claude # Chat history is saved between sessions. # Use Alt+Enter for multiline input in chat mode. # In chat, type /edit to open your $EDITOR and compose a message. # Press Ctrl+O in chat to open a buffer for editing your prompt directly. rullm "Tell me a story" rullm --model gpt4 "Explain quantum computing in detail" rullm chat --model claude # Disable streaming for buffered output rullm --no-streaming "Write a poem about the ocean" rullm chat --no-streaming --model claude # Set up your API keys rullm keys set openai export OPENAI_API_KEY="your-key-here"You can pipe files or stdin into rullm and optionally add a query string. The CLI will merge both, making it easy to work with code or text files:
# Just pipe a file (stdin only) cat foo.py | rullm # Pipe a file and add a query (stdin + arg) cat foo.py | rullm "explain this code"You can pass a system prompt directly to the model using the --system argument. This lets you customize the LLM's behavior for a single request:
rullm --system "You are a helpful assistant." "Summarize this text"# List available models (shows only chat models, with your aliases) rullm models list # Update model list rullm models update # Manage aliases rullm alias list rullm alias add my-fast "openai:gpt-3.5-turbo" rullm alias show claude # API key management rullm keys set openai rullm keys list# Use a template ({{input}} is replaced by your query) rullm -t my-template "input text"Templates are stored as TOML files in ~/.config/rullm/templates/ (or your system's config directory):
name = "code-review" description = "Template for code review requests" # You can include multi-line prompts using TOML triple-quoted strings: system_prompt = """ You are a senior Rust engineer. Provide a thorough review with the following structure: 1. Summary 2. Strengths 3. Weaknesses 4. Suggestions """ user_prompt = "Please review this code: {{input}}"{{input}}– Automatically filled with the user's query text.
Model aliases are now user-defined. Use rullm alias add <alias> <provider:model> to create your own shortcuts. Use rullm alias list to see your aliases.
To enable shell completion, generate the completion script for your shell:
# fish source (COMPLETE=fish ./target/debug/rullm | psub) # bash source <(COMPLETE=bash ./target/debug/rullm) # zsh source <(COMPLETE=zsh ./target/debug/rullm)