Manual Test Plan

These tests require the user’s real machine, at least one configured text model provider, and installed ComfyUI environment.

First Launch

1. Install/open the packaged Museboard.app.
2. Confirm the app opens without terminal.
3. Confirm no browser demo text appears in the desktop app.
4. Switch language to English, close settings, reopen settings.
5. Confirm language setting persists after app restart.

Text Model Providers

1. Choose LM Studio, enter endpoint http://localhost:1234/v1, and enter a loaded model name.
2. Click 随机导图.
3. Select a node and click 生成说明.
4. Select a node and click 生成下一级.
5. Confirm generated text is visual, concrete, and in the selected UI language.
6. Switch to Ollama, enter endpoint http://localhost:11434, and enter a pulled model name.
7. Generate one node description.
8. If OpenAI or DeepSeek keys are available, switch to that provider, enter endpoint, API key, and model name, then generate one brief.
9. Close and reopen Museboard. Confirm provider, endpoint, model, API key, temperature, and language persist.

ComfyUI

1. Ensure ComfyUI is not running.
2. Open 设置 -> ComfyUI 连接向导.
3. Set ComfyUI folder and launch command.
4. Click 启动并连接.
5. Confirm ComfyUI starts automatically.
6. Click Flux 推荐预设.
7. Click 生成测试图.
8. Confirm a test image appears.
9. Close settings.
10. Click main 生成图片.
11. Confirm the prompt is related to the current brief.
12. Generate an image and save it to the root node.

Project Save/Restore

1. Create a project.
2. Add at least 20 images across several child nodes.
3. Add rectangle, arrow, text, and freehand annotations.
4. Use parent node view to rearrange child images.
5. Save the project to a new folder.
6. Quit the app.
7. Reopen the project folder.
8. Confirm nodes, images, layouts, annotations, settings, and generated images remain correct.

Canvas

1. Drag local images from Finder into a child node.
2. Drag an image link from a browser into a node.
3. Use selection box to move several images and annotations together.
4. Use scroll wheel or trackpad scroll to pan the canvas.
5. Use Cmd + scroll or trackpad pinch to zoom around the pointer.
6. Use Cmd + =, Cmd + -, and Cmd + 0 to zoom in, zoom out, and reset view.
7. Hold Space and drag on the canvas. Confirm the canvas pans and images are not moved.
8. Tap Space with one image selected. Confirm the image fits the view; tap Space again and confirm the previous view returns.
9. Tap Space with no image selected. Confirm all visible images fit the view.
10. Double-click an image to make it a 75 x 75 thumbnail, then double-click again to restore it.
11. Press Tab. Confirm clean canvas mode hides extra UI and leaves the canvas/tools visible.
12. Press Tab again. Confirm the hidden UI returns.
13. Confirm Tab still moves focus normally while editing text fields or text areas.

Failure Cases

1. Stop ComfyUI and click 测试连接.
2. Confirm the error tells the user what to do.
3. Import non-API workflow JSON.
4. Confirm the workflow error is understandable.
5. Temporarily use a wrong model name in the selected text model provider.
6. Confirm Museboard does not crash.