Configuration

⚙️ Configuration

Learn how to configure Maestro V2 for your needs.

---

Model Presets

Maestro V2 supports 4 model presets:

Preset	Best For	Models Used
Basic	Quick tests	llama3:8b (all agents)
Standard	Balanced	llama3:70b + codellama:13b
Advanced	Production	llama3:70b + codellama:34b
Cloud	Maximum power	GPT-4o + Claude 3.5

Changing Presets

In the UI, use the model preset dropdown in the header.

Via environment:

export MODEL_PRESET=advanced
python app.py

---

Environment Variables

Variable	Default	Description
OLLAMA_API_URL	http://localhost:11434/api/generate	Ollama API endpoint
OPENAI_API_KEY	(none)	OpenAI API key for cloud
ANTHROPIC_API_KEY	(none)	Anthropic API key for Claude
MODEL_PRESET	basic	Default model preset

---

Agent Temperatures

Each agent has a configured temperature for creativity:

Agent	Temperature	Behavior
Developer	0.3	Consistent, precise
Security	0.3	Consistent, precise
QA	0.4	Balanced
Research	0.5	Balanced
Documentation	0.5	Balanced
UI/UX	0.6	Creative
Refiner	0.6	Creative
Orchestrator	0.7	Strategic

---

Project Storage

Projects are saved to:

~/.maestro/projects/<project-id>/
├── config.json
├── logs/
│   ├── Orchestrator.md
│   ├── Developer.md
│   └── ...
└── output/
    └── final_output.md

---

Database

SQLite database is stored at:

maestro-agents/maestro.db

Contains:

• Project history
• Workflow records
• Agent execution metrics
• Memory entries

---

← Tools | Home →