Before practicing, learn the shortcuts and check basic settings.
Before a written test or online assessment, read Written-test platform compatibility. Standalone test clients require Windows, administrator launch and a tested dual-device link.
Screenshot mode
- Full-screen screenshot: covers the entire screen.
- Partial screenshot: manually select an area for higher precision.
Questions can include code, SQL, mathematics, aptitude-test choices or written responses. The solution follows the question's requirements. For coding problems, include the function signature, input/output requirements, examples and constraints.
Screenshot-solving model controls this flow; Programming language applies when code is needed. Answer question and Voice solve are separate voice-panel actions.
Practice a complete solution
INTERACTIVE WALKTHROUGH
Try it: from screenshot to solution
Click the numbered markers to see what happens next in the client, or use the tutorial controls below.
Open the question and click Screenshot

Show the question, input/output requirements and examples on screen. Click the highlighted Screenshot control. In partial screenshot mode, then select the question area.
What you should see: A thumbnail appears above the toolbar, ready to submit with Solve.
For a long question, scroll the question and take additional screenshots before solving.
Default shortcutMac ⌘ + HWindows Ctrl + H
Show all illustrated steps
1. Open the question and click Screenshot
Show the question, input/output requirements and examples on screen. Click the highlighted Screenshot control. In partial screenshot mode, then select the question area.

A thumbnail appears above the toolbar, ready to submit with Solve.
For a long question, scroll the question and take additional screenshots before solving.
2. Check your screenshots, then click Solve
Check the screenshot queue above the toolbar. Once the question and examples are captured, click the highlighted Solve control.

The answer panel opens and starts processing your screenshots.
Delete an incorrect screenshot and capture it again. If Solve is unavailable, check that a screenshot is queued and the previous generation has finished.
3. Wait for the answer to appear
No further click is needed in the client. A processing message appears before the answer starts arriving.

Question analysis, an approach or code appears in the answer panel.
Timing depends on the model and network. Avoid submitting repeatedly. This tutorial uses a fixed practice question and answer.
4. Click Next page inside the answer panel
Read the visible answer, then click Next page above the answer panel, or use the corresponding paging shortcut.

The same answer scrolls down to reveal more content. It does not generate a new answer.
Previous page scrolls back up. The default shortcuts use comma and period without Shift; the interface displays these as < and >.
5. Click Screenshot again for an error or new condition
Show your current code, error, failing case or new condition in the editor or question page, then click Screenshot. Select the relevant area for a partial capture, and take more screenshots when needed.

The new screenshots are queued for this question while the existing answer remains available.
Feedback is provided through screenshots. Include actual and expected output with the related case. This tutorial uses fixed practice feedback.
6. Check the feedback, then click Continue this question
Check the additional screenshots and click Continue this question. Wait for a complete revised answer based on the question, existing solution and new feedback.

The revised answer appears. Add further screenshots and submit again when needed.
Continue this question needs an additional screenshot. It processes new feedback; Next page only scrolls the current answer.
7. Capture the next question directly
Check and save the revised answer. Show the next question on screen and click Screenshot.

The next question is queued while the current answer remains visible. Solve starts an independent question.
For another revision, add screenshots and continue this question. If local saving failed, use Retry saving before starting another question.
8. Use Solve for the newly captured question
Check the new screenshot, then click Solve. Use Continue this question only for changes to the existing question.

You have completed: capture → solve → read more → add feedback → continue this question → capture the next question.
Replay below only restarts this website tutorial. Perform the corresponding actions in the desktop client when using the tool.
Continue with errors or new conditions
- Keep the current answer and show your code, error messages, failing cases, actual and expected output, or new conditions in your editor or question page.
- Click Screenshot or use the default
⌘ / Ctrl + H. Select the relevant area in partial capture mode. Take multiple screenshots when feedback spans several areas. - Check the additional screenshots, then click Continue this question or use the default
⌘ / Ctrl + D. The model uses the existing question and solution together with the new screenshots. - Read the complete revised answer. Add more screenshots for further feedback. For a different question, capture it and click Solve or use the default
⌘ / Ctrl + Enterto start a separate question. Use Start Over or⌘ / Ctrl + Rwhen you want to clear the current question and pending screenshots.
Provide feedback through screenshots while continuing to work in your editor or question page. Previous page and Next page only scroll the current answer.
Saving and retrying
Completed answers and their screenshots are saved to local exam records. Continuing the same question updates the complete answer and archives the additional screenshots.
If local saving fails and screenshots are retained, click Retry saving before starting another question. This saves the already generated answer without calling the model again. For a generation failure, check the network and model configuration before submitting the screenshots again.
Related docs
Keyboard shortcuts
Learn default key combinations and customization to improve operation efficiency.
Custom API and models
Continue configuring API, models, and multimodal capabilities.
Practical preparation
Complete stealth, voice, and rehearsal checks before interviews.
Written-test compatibility
Check capture and phone requirements by system and test platform.