← Prev: Project Structure | Home
| Problem | Cause | Solution |
|---|---|---|
| Emojis displayed as outlines in Flatpak | Flatpak sandbox blocks system emoji fonts | Use native package or AppImage |
| FC12/FC13 disabled for Ollama | Too complex for local models | Use Claude (Anthropic) or GPT-4o |
| API keys stored unencrypted | FreeCAD settings have no encryption | Do not use production keys |
| Large files (>2,000 lines) | AI context window is limited | Load only relevant sections into the input field |
| Ollama not found | Service is not running | Run ollama serve in the terminal |
| Ollama produces poor FreeCAD code | General model without code focus | Install ollama pull qwen2.5-coder:7b |
| Aborting a run only stops between Python lines | A running FreeCAD C++ operation cannot be interrupted | Wait for the current operation; the abort takes effect at the next line |
| Stopping an AI request waits for the next chunk | The stream breaks at the next received line, not mid-socket | A fully stalled provider still ends via the 120 s timeout |
If you want, I can split or expand any of these into separate troubleshooting pages.