Windows Click to Do is a screen overlay: Win+click captures what is on screen, and Microsoft's on-device models act on the rendered text. It needs a Copilot+ PC. OpenCues works in the text field itself, as you type, with no special hardware and no overlay detour.
What Click to Do does well
It acts on anything visible, even text that is not editable, and it ships free with Windows on qualifying machines.
Where Click to Do struggles
It operates on a screenshot. The OCR overlay acts on a picture of your text, not the text control itself, so it can read anything visible but write back nothing in place: results land in panels and copies. It also requires Copilot+ NPU hardware, which excludes most Windows machines in use today.
How OpenCues differs
OpenCues is inline rather than overlay: the _ you type is the interface, and the result lands in your text. It runs on any Windows machine through Chrome and WSL terminals, needs no NPU, and is open source and model agnostic, from cloud providers to your existing AI subscription to local Ollama.
Click to Do output replaces what it touches; OpenCues edits land revertable, the original is one cycle away, per word.
OpenCues is also extensible by design: cues and blanks are markdown files you write yourself, and blanks can bind shell scripts or runtime classes, all under an open standard anyone can implement. See write your own cue and write your own blank.
Related: Full comparison · AI in any text field · Windows support