Frequently Asked Questions¶

Cost & Billing¶

Q: Why don't my tracked costs match my API provider dashboard?¶

A: This is a known issue, particularly with API aggregators like OpenRouter. MAESTRO calculates costs based on advertised pricing, but actual charges can vary because:

Aggregators route to different backend providers with varying costs
Dynamic routing optimizes for speed/availability, not just price
Some providers count hidden tokens not reported in their API

Your tracked costs may typically be 40-60% of actual dashboard charges, especially with providers like Openrouter, that may route your calls to further providers with differing rates. This is not a bug in MAESTRO - we calculate correctly based on advertised rates. See Cost Tracking Discrepancies for details and workarounds.

Q: How can I reduce my API costs?¶

A: Several strategies can help:

Use cheaper models for Fast/Mid tiers
Reduce research parameters in Settings → Research
Use local models for zero API costs
Monitor actual dashboard charges, not just tracked costs

Models & Providers¶

Q: Which AI provider should I use?¶

A: It depends on your needs:

OpenAI: Most consistent pricing and reliability
OpenRouter: Access to 100+ models, but pricing can be inconsistent
Local Models: Zero API costs, but requires GPU/CPU resources

Q: Can I use local LLMs?¶

A: Yes! MAESTRO supports any OpenAI-compatible endpoint. See our Local LLM Deployment Guide for setup instructions.

Q: How do I configure Azure OpenAI?¶

A: Azure OpenAI requires specific URL formatting:

Select "Custom Provider"
Base URL: https://your-resource.openai.azure.com/openai/v1/
Must end with /openai/v1/ (not /openai/deployments/)
Enable "Manual Model Entry" toggle
Enter your Azure deployment names (not model names)

Note: Different providers may require specific URL path suffixes. Always verify the correct format: - Azure OpenAI: /openai/v1/ - Most others: /v1/

See AI Provider Configuration for details.

Common Issues¶

Q: Why are my responses slow?¶

A: Check these common causes:

Using large/slow models (try faster models like gpt-4o-mini)
High context sizes in Research settings
Network latency to API provider
Rate limiting from provider

Q: Why do I get "context too large" errors?¶

A: Reduce these settings in Settings → Research:

writing_agent_max_context_chars
main_research_doc_results
main_research_web_results

More Help¶

For detailed troubleshooting, see: - AI Model Troubleshooting - Database Issues - Installation Problems