CLI Usage¶

The Savanty command-line interface provides quick access to the optimization solver from your terminal.

Basic Usage¶

Bash

savanty -p "Your problem description"

Commands and Options¶

Solve a Problem¶

Bash

savanty -p "Schedule 3 workers for 2 shifts"
savanty --problem "Assign tasks to machines"

Start Web Interface¶

Bash

savanty --web                    # Default port 8000
savanty --web --port 3000        # Custom port
savanty -w -p 3000               # Short form

Run as MCP Server¶

For Model Context Protocol integration:

Bash

savanty --mcp
savanty -m

Show Version¶

Bash

savanty --version

Show Help¶

Bash

savanty --help

Interactive Mode¶

When Savanty needs more information, it enters interactive mode:

Text Only

$ savanty -p "Schedule nurses for shifts"

I need more information to solve this problem:
1. How many nurses are available?
2. What shifts exist?
3. How many days should the schedule cover?

Please provide the missing information: 4 nurses, morning/afternoon/night shifts, 7 days

The solver continues with your additional information.

Exit Codes¶

Code	Meaning
0	Success - solution found
1	Error - problem couldn't be solved

Examples¶

Simple Scheduling¶

Bash

savanty -p "Schedule Alice and Bob for morning and evening shifts on Monday and Tuesday. Each shift needs one person, no one works both shifts on the same day."

With Pipe¶

Bash

echo "Assign 3 tasks to 2 workers" | savanty -p -

In a Script¶

Bash

#!/bin/bash
PROBLEM="Schedule maintenance for 5 machines over 3 days"
RESULT=$(savanty -p "$PROBLEM")
echo "Solution: $RESULT"

Environment Variables¶

Set these before running:

Bash

export OPENAI_API_KEY=your_key_here
export SAVANTY_LLM_MODEL=openai/gpt-4o
export SAVANTY_PORT=8000
export SAVANTY_LOG_LEVEL=INFO

Common Issues¶

"OPENAI_API_KEY not configured"¶

Set your API key:

Bash

export OPENAI_API_KEY=your_key_here

Problem Timeout¶

For complex problems, increase the timeout:

Bash

export SAVANTY_SOLVE_TIMEOUT=300  # 5 minutes
savanty -p "Complex scheduling problem..."

"Not suitable for ASP"¶

Some problems are better solved with other tools. Savanty will suggest alternatives:

Text Only

This problem is better suited for a different tool.
Reason: This appears to be a continuous optimization problem.
Suggested tool: scipy (pip install scipy)

Tips¶

Quote your problem to preserve spaces and special characters
Use multi-line input for complex problems by using quotes
Check logs with SAVANTY_LOG_LEVEL=DEBUG for troubleshooting