Product

Auto-capture

Click through your workflow and Knovvy writes the steps, grabs the screenshots, and highlights exactly what you clicked. Built from real browser events, so it is exact rather than inferred.

Record your first guide

Documentation that writes itself

🖱️
Clicks to stepsEach action becomes a numbered instruction named after the element you actually touched.
🖼️
Screenshots + highlightsEvery step gets an image with the click target highlighted on its real bounding box.
🕶️
Auto-blurPassword, one-time-code and card fields are redacted before the screenshot is stored. No AI.

Install the Knovvy extension

Guide capture runs in your browser. Add the free Chrome extension, click through your workflow, and the steps, screenshots and highlights are written for you.

Add to Chrome, free

How it works

1

Record

Hit record, then click through your workflow the way you normally would.

2

Auto-generate

Knovvy writes the steps, screenshots, captions and chapters, and keeps the video.

3

Share

Send a link, embed it, or export to PDF, HTML, or Markdown in one click.

How it actually works

Events, not pixels

There are two ways to turn a recording into a guide: watch the video and infer what happened, or be present while it happens and record what did. Knovvy does the second.

  • The instruction comes from the element you interacted with, using its accessible name first, then its visible text, then its placeholder or title.
  • The highlight sits on the real bounding box, not an estimate.
  • Typing collapses into one step per field when the value is committed, rather than one per keystroke.
  • Each step records the page title, hostname, and URL, which is how a guide names itself after the tool and the task.

The result is identical every run, which is a property inference cannot offer at any model size. Longer argument: deterministic capture beats AI guessing.

Capture follows you across tabs and sites

A workflow that spans three tools is still one guide. When capture crosses to a new host it inserts a "Go to example.com" step at the boundary, so the reader knows to switch too.

Screenshots are serialized with a minimum gap, so a burst of fast clicks still produces one image per step instead of several steps sharing a frame.

What happens when a page cannot be captured

Browser-internal pages (chrome:// and the Web Store) cannot be screenshotted by any extension. Rather than storing a blank image or dropping the step, Knovvy records the step and labels it in the editor, and the on-page badge warns you while you are on such a page.

Recovering from a mistake mid-capture
  • Pause stops recording without ending the session. Nothing is captured while paused.
  • Undo removes the last step and its screenshot. Press it repeatedly to walk back several.
  • Discard throws the session away without creating a guide.

All three live on the draggable on-page badge, with the toolbar popup as a fallback. Detail: pause, undo, and discard mid-capture.

Common questions

Does auto-capture need AI?

No. Steps, screenshots, click highlights, and sensitive-field blurring are all deterministic, derived from real browser events. You can capture guides on the free tier forever without touching a model.

Does the extension record when I am not capturing?

No. The content script loads on pages you visit, but every listener is gated on an active capture session. Until you press record, no step is created, no screenshot is taken, and nothing is sent to our servers.

Can it capture a workflow that spans several apps?

Yes. Capture follows you across tabs and sites, and inserts a "Go to example.com" step when you cross to a new host so the guide reads correctly.

Ready to capture the know-how?