Commands Overview¶

ScoutML provides a comprehensive set of commands organized into logical categories. Each command is designed to help you efficiently navigate and understand ML research.

Command Categories¶

🔍 Search Commands¶

Find papers using various search strategies:

search - Natural language semantic search
method-search - Find papers using specific methods
dataset-search - Find papers using specific datasets

📊 Analysis Commands¶

Analyze and compare papers:

paper - Get detailed information about a paper
compare - AI-powered comparison of multiple papers
similar - Find papers similar to a given paper

📝 Research Synthesis¶

Generate comprehensive research summaries:

review - Generate AI-synthesized literature reviews

🤖 Intelligent Agents¶

AI-powered research assistance:

agent implement - Generate implementation guides
agent critique - Get research critique and analysis
agent solve-limitations - Find solutions to paper limitations
agent design-experiment - Design validation experiments

💡 Research Intelligence¶

Analyze research trends and patterns:

insights reproducibility - Reproducibility analysis
insights funding - Funding source analysis

Common Options¶

Most commands support these common options:

Option	Description
`--output FORMAT`	Output format: `table`, `json`, `csv`, `markdown`, `rich`
`--export PATH`	Export results to a file
`--help`	Show command help and options

Output Formats¶

ScoutML supports multiple output formats:

table (default) - Rich terminal tables with colors
json - Structured JSON for programmatic use
csv - Comma-separated values for spreadsheets
markdown - Formatted markdown for documentation
rich - Enhanced terminal output with panels

Using Help¶

Get help for any command:

# General help
scoutml --help

# Command-specific help
scoutml search --help
scoutml agent implement --help

Command Chaining¶

ScoutML commands can be chained together for powerful workflows:

# Search → Export IDs → Compare
scoutml search "BERT variants" --output json | \
  jq -r '.[] | .arxiv_id' | \
  head -3 | \
  xargs scoutml compare

Error Handling¶

ScoutML provides clear error messages:

Missing API Key: Instructions on configuration
Invalid Parameters: Helpful suggestions
No Results: Tips for broadening search
Rate Limits: Retry timing information

Best Practices¶

Start broad, then narrow: Use filters to refine results
Use appropriate output formats: JSON for scripts, table for reading
Export important results: Use --export to save findings
Combine commands: Build workflows for complex tasks
Check help often: Commands have many useful options

Next Steps¶

Explore specific command categories:

Start with search commands to find papers
Use analysis commands to understand papers
Try AI agents for implementation help
Discover research insights for trends