What is --output-format in Claude Code
The --output-format flag controls how Claude Code structures its responses in non-interactive mode, enabling JSON output for programmatic processing or formatted text for human consumption.
How to Use It
Use --output-format with --print mode to specify response structure: text for human reading, json for programmatic processing, or stream-json for real-time data handling.
claude -p "Analyze code quality" --output-format json
claude -p "Generate documentation" --output-format text
claude -p "Large analysis" --output-format stream-json
Why Use --output-format
The --output-format flag enables Claude Code integration with data processing pipelines, automation scripts, and tools requiring structured data input.
Benefits:
- Programmatic Processing - JSON output for scripts and automation workflows
- Data Integration - Structured responses for database storage and API integration
- Pipeline Compatibility - Stream processing for large responses and real-time handling
- Parsing Efficiency - Machine-readable formats for complex data extraction
- Tool Interoperability - Compatible with jq, Python scripts, and CI/CD systems
Format Options
Text - Human-readable plain text (default format).
JSON - Structured data with metadata, analysis results, and processing information.
Stream-JSON - Real-time streaming JSON for large responses and progressive processing.
Common Use Cases
CI/CD Integration - Extract specific metrics from Claude Code analysis for build pipelines.
Script Automation - Process JSON responses with jq, Python, or other data processing tools.
Data Analysis - Store structured Claude Code outputs in databases or generate reports.
--output-format json is crucial for CI/CD integration and automation workflows. It enables precise data extraction and programmatic processing of Claude Code analysis.
See Also: Print Mode|Max Turns|Configuration