Calculate:
arcsin(−1/2)
An answer to your task right from your screen
The app takes a screenshot, a neural network extracts text and images, then provides an answer with a solution.
- Solve a test
- Solve a problem
- Translate text
- Analyse code
- Answers in a few seconds
- Web control from another device
- Invisible during screen sharing (on PC: overlay, remote control)
- Invisible to the target page (on PC: overlay, remote control)
Let's see what you've got there...
1 click = 3 steps to the answer
Press a button or hotkey: ChiMate captures a screenshot, extracts task text with OCR, recognises related images and prepares an answer. Response time depends on the selected model.
Step 01
Screenshot capture
Take a screenshot of any window using a hotkey or by pressing the button on the special translucent control panel (the "Overlay"). No context switching and no loss of focus.
Step 02
Text and image recognition
ChiMate extracts the text of the task with OCR, preserving its structure. It also recognises related images, including drawings, diagrams and graphs, for further analysis.
Step 03
AI analysis
The model takes the task's text and images into account to produce a structured answer: an explanation, translation or solution.
One assistant. Four ways to work
Open the main window, keep an answer over your task, capture a long page or control ChiMate from another device. Choose a mode to see it in action. Learn more.
ChiMate.aiOnline570 CRThis is a Chinese word; the other options are Japanese and Korean.
Your task and answer side by side
Select the window with your task and take a screenshot. ChiMate shows the answer with an explanation — next to your work or on a second monitor.
- Capture a window, region or screen
- Choose the model and task type
- Keep your answers in one history
- For stealth operation, use other control methods: overlay, remote control, or keyboard shortcuts.
Support for many languages and work modes
Choose different languages and work modes, or leave it in automatic mode.
Multiple AI tiers are available — from fast and economical models to the most capable ones with the highest level of reasoning.
Area of responsibility — the whole world
Main modes:
- Tests — tasks with multiple-choice answers
- Translation — foreign-language tasks that need to be translated into your native language
- Code — analysing or writing source code
Three AI model tiers to choose from — from speed to maximum intelligence
Different requests suit different AI tiers. Pick the balance between speed, price and quality that matches the task at hand.
Tier
Simple
Fast and economical for simple scenarios.
- Price
- 1 credit / request
- Accuracy
- ≈ 70%
- Speed
- Fast
Best for: simple questions, tasks, translations
Not for: coding, complex tasks, high-stakes tasks
Tier
Advanced
Versatile balance of quality and cost.
- Price
- 2 credits / request
- Accuracy
- ≈ 85%
- Speed
- Average
Best for: most everyday tasks, coding, tests
Not for: high-stakes tasks that require the highest possible accuracy
Tier
Top
Maximum accuracy for critical tasks.
- Price
- 3 credits / request
- Accuracy
- ≈ 95%
- Speed
- Slow
Best for: the most important tasks, exams, complex code
Not for: bulk simple requests — would be unnecessarily expensive
Frequently asked questions
Answers to the most common questions. Can't find yours — drop us a line.
What are credits?
Credits are spent on each analysis request. The cost depends on the chosen model tier: Simple — 1 credit, Advanced — 2, Top — 3. Your plan determines how many credits you receive each period.
Which model tiers are available?
Three tiers are available — Simple, Advanced and Top. Higher tiers use more capable models and deliver up to 95% accuracy on complex tasks.
Will other apps or websites notice the overlay and screenshots?
The Windows app can exclude its overlay from compatible screen captures when capture protection is enabled. The browser assistant does not guarantee this: its window may appear when sharing an entire screen.
Which operating systems are supported?
The desktop app supports Windows 10/11 (x64). The browser assistant works in supported desktop browsers, including current Chrome and Edge. Screen capture and floating-window support depend on the browser.
What languages do OCR and analysis support?
OCR recognises virtually any language supported by the selected model. Analysis answers automatically adapt to the interface language or the source task language.
Does ChiMate recognise images in a task?
Yes. ChiMate extracts task text with OCR and recognises related drawings, diagrams and graphs. The model takes both text and images into account when preparing an answer. Capture the entire task and its related images.
How long does it take to get an answer?
It depends on the selected model tier, the complexity of the task and the number of questions in it. Approximate response times: Simple — 2–20 seconds, Advanced — 5–40 seconds, Top — up to 60 seconds.
What if the task is too long to fit on the screen?
For these cases, CDP mode can capture the entire page, including content beyond the visible area.
Can I get a refund if the app does not suit me?
You can request a refund if you have used no more than 30 credits. Requests are reviewed within 3 business days.
Download now and try it for free on your own tasks!
No installation - download and run. No registration - sign in with Google.