Performance Optimization

General Performance

• Close unused editor tabs
• Disable unnecessary extensions
• Use files.exclude to hide large directories (e.g., node_modules)

AI Response Speed

• Gemini — Response speed depends on model and API load. Use gemini-pro for faster responses.
• Ollama — Use smaller models (codellama:7b) on machines with limited resources. GPU acceleration significantly improves speed.

Large Projects

For projects with 10,000+ files:

• Exclude build artifacts and dependencies from the workspace
• Use workspace-scoped settings to limit agent context

Related

• Common Issues
• Settings Reference