CLI Reference

Commands

flyto-ai (default)

Start an interactive chat session.

flyto-ai

flyto-ai chat

One-shot message execution.

# Execute a task
flyto-ai chat "scrape the top 10 posts from Hacker News"

# YAML-only mode (no execution)
flyto-ai chat "scrape example.com" --plan

# Use specific provider
flyto-ai chat "..." -p ollama

# Use specific model
flyto-ai chat "..." -p openai -m gpt-4o-mini

# Send result to webhook
flyto-ai chat "..." --webhook https://hook.site/xxx

flyto-ai code

Claude Code Agent for code tasks.

# Basic code fix
flyto-ai code "fix the login page" --dir ./project

# With screenshot verification
flyto-ai code "fix the login page" --dir ./project --verify screenshot

# With visual comparison to reference
flyto-ai code "match this design" --dir ./project --reference ./figma.png

flyto-ai serve

Start HTTP server for external triggers.

flyto-ai serve --port 8080

flyto-ai blueprints

Manage learned workflow blueprints.

# List all blueprints
flyto-ai blueprints

# Export as YAML
flyto-ai blueprints --export

flyto-ai mcp

Start as an MCP server (JSON-RPC over STDIO).

flyto-ai mcp

flyto-ai version

Show version and dependency information.

flyto-ai version

flyto-ai prompt-lab

Prompt optimization tools.

# Evaluate prompt against test cases
flyto-ai prompt-lab eval --cases cases.yaml --rubric rubric.yaml

# Evolve prompt using genetic algorithm
flyto-ai prompt-lab evolve --generations 5 --population 10

Global Options

Option	Description
-p, --provider	LLM provider (openai, anthropic, ollama)
-m, --model	Model name
--plan	YAML-only mode, no execution
--webhook	POST results to URL
--dir	Working directory for code tasks