A CLI tool for querying LLMs

go install github.com/mkozhukh/qq@latest

Set your API key and preferred model using environment variables:

export QQ_KEY="your-api-key"
export QQ_MODEL="provider/model-name"
export QQ_PROMPT="Optional system prompt"

Supported providers and example models:

• openai/gpt-4.1-mini
• anthropic/claude-4-0-sonnet
• google/gemini-2.5-flash
• openrouter/moonshotai/kimi-k2

qq "What is the capital of France?"

When run without arguments, qq enters interactive mode

qq
# Enter your message in the form that appears

cat error.log | qq "What's causing these errors?"

git diff | qq "Summarize these changes"

Add context with system messages:

qq --system-message "You are a senior DevOps engineer" "How do I set up a CI/CD pipeline?"

Use echo-templates for advanced prompting:

# Use inline template
qq --prompt "You are {{role}}. {{user_query}}" "Explain Docker"

# Use template file
qq --prompt-file "templates/code-review.md" "Review this code"

# Use template folder
qq --prompt-folder "./templates" --prompt-file "analyze.md" "Analyze this log"

Use --markdown flag to format output as markdown:

qq --markdown "Explain the TCP/IP model"

• QQ_MODEL: Default model (e.g., openai/gpt-4.1)
• QQ_KEY: API key for the provider
• QQ_PROMPT: Default system message

if default variables are not defined, code fallback to the ones from mkozhukh/echo

• ECHO_MODEL - for model
• ECHO_KEY - for API key

if api key is not found, it goes deeper

• GEMINI_API_KEY
• OPENAI_API_KEY
• ANTHROPIC_API_KEY
• OPENROUTER_API_KEY

• --model: Override the model for this query
• --key: Override the API key for this query
• --system-message: Add a system message for context
• --prompt: Prompt template (echo-templates format)
• --prompt-file: File containing prompt template
• --prompt-folder: Folder containing prompt template files
• --markdown: Format output as markdown
• --debug: Show debug information
• --help: Show usage information

MIT License - see LICENSE file for details