Imageanalysis.

The img: command captures a region of your screen and asks the AI to read it. One requirement, four typical uses.

llama3.2-vision:11bRequiredOCRbuilt inOptional

PLACEHOLDER

Captured region

The area Typilot sends to the vision model + OCR engine.

OCR + VISION OUTPUT

# text extracted

"Connection failed - error code 0x80072EE7"

# model interpretation

Windows network error: DNS failed to resolve the target host. Most likely cause: VPN intercepting traffic or stale DNS cache. Try ipconfig /flushdns.

One dependency, free, downloaded once. OCR is already built in.

Vision-capable model. e.g. llama3.2-vision:11b. Download from the Typilot Models page.
OCR. Ships inside Typilot and reads text from the captured region. Nothing to install.

Analyze screenshots. Capture and ask. Useful for complex UIs or screens you don't recognize.
img: describe what's happening in this screenshot
Extract text from images. OCR pulls the text, AI formats and summarizes.
img: extract and format the table data
Decipher error dialogs. Plain-language explanations of cryptic system messages.
img: what does this error mean?
Inventory UI elements. Helpful when documenting interfaces or learning a new app.
img: list the buttons visible on this screen