How many screenshots can I analyze with 100 free credits?
During the public beta, new InvisiCopilot accounts receive 100 free signup credits to try AI written-question assistance without buying a plan first. Signup credits have long-term validity and no automatic refill.
Each example below uses all 100 credits for one capture mode. The allowances cannot be combined. Interview answers and captures share the balance; a Quick-to-Deep fallback costs 8 credits in total.
| Using only this feature | Credit cost | Allowance with 100 free credits |
|---|---|---|
| Quick Capture without Deep fallback | 4 credits per capture | 25 captures |
| Deep Capture | 8 credits per capture | 12 captures, with 4 credits left |
Free access and plans: Free trial details
Choose capture mode for the question
Both modes can make mistakes. Check the question, inputs, outputs, and diagram labels before relying on hints. When recognition fails, no question is found, or content is unclear, the client displays the reason and refunds the actual capture charge.
| Question content | Suggested mode | Cost on credit plans |
|---|---|---|
| Mostly text: coding, multiple-choice, fill-in, and true/false | Quick Capture | 4 credits per capture |
| Geometry, charts, spatial or connection relationships | Deep Capture reads text and graphics directly | 8 credits per capture |
| Quick Capture needs original-image analysis | Fallback to Deep Capture | 4 additional credits; 8 in total |
Written practice without starting audio
- Install and sign in, then grant system recording permissions as prompted. Mac requires macOS 13+ on Apple Silicon; Windows 10 and 11 are supported. AI capture requires an internet connection.
- End any interview or Audio Check, then choose OA mode from the menu. It keeps the capture window without interview audio or an interview timer. Capture still has a per-use rate.
- Display the complete question, inputs, outputs, and necessary diagrams. Choose Quick or Deep Capture.
- Check the recognized question and read the explanation and code. Run the code in your own practice environment and check samples and edge cases.
- Capture again for another question. Up to three captures across both modes can run at once, with independent result pages.
Setup details: Capture controls and shortcuts · OA mode
Coding practice example: check the key assumptions
For Word Ladder, take hit as the start, cog as the target, and hot, dot, dog, lot, log, cog as the dictionary. One shortest path is hit → hot → dot → dog → cog, with length 5. After capture, check whether the explanation uses level-order search, maintains the visited set correctly, and counts words rather than transitions.
Also check cases where the target is missing, dictionary entries repeat, or no path exists. Run and verify generated code yourself. This is a practice verification example; it does not mean the client runs your code or establish an accuracy measurement.
Navigate results and manage history
Use the previous/next controls or Shift + Left/Right to switch result pages. Shift + Up/Down scrolls the current answer. A completed background task does not replace the page you are reading.
Cloud saving is on by default. Standalone captures can save results and original model responses from each stage, including partial outputs received before a failed request. Disable future saving in Interview Management and delete screen records on the history page. Authorized administrators may access saved records for support.
Data details: Privacy policy
How much does written-question assistance cost?
During the public beta, registration includes 100 credits. On credit plans, OA mode starts no per-minute interview billing. Quick Capture costs 4 credits and Deep Capture costs 8. Five successful Quick Captures without fallback use 20 credits; two Deep Captures use 16.
Active Monthly Sprint plans waive capture credit charges; concurrency limits still apply. Packs are one-time purchases with no automatic renewal. See the plans page for current prices and validity periods.
Compare plans: Prices and entitlements
Need live interview transcription as well?
Screen capture and interview audio are separate features. Leave OA mode and explicitly start an interview when you want spoken practice; leaving the mode does not start the timer. See the interview assistant page for bilingual transcription, resume context, and review settings.
Explore the other workflow: AI interview assistant
Use cases and limits
Use for practice or where the organizer explicitly permits AI and external tools. Statutory national examinations and their interview stages are unsupported. Dedicated exam clients may block screenshots; capture failure does not establish a bypass, and no visible alert does not establish an absence of monitoring.
AI answers and code can be wrong and do not guarantee assessment or interview outcomes. Before buying, test your operating system, question types, and workflow with practice material you are permitted to use.
Scope of use: Terms of service