Walkthroughs · Automated

Product walkthroughs, generated from words.

Describe the walk. Tuti Tuti takes it, on your real app, with a voice and a cursor, and hands you the video.

Generate a walkthroughSee pricing

What makes a walkthrough different from a recording

A recording is a memory of one run. A walkthrough made by Tuti Tuti is a program that runs again: each sentence is pinned to the element it names, the cursor follows the same path, and the narration lands on the same words. Change the app and run it again; change the words and run it again. The video is always a fresh performance.

How the walk is performed

The mechanics, since developers ask.

01

Rules decide act, show, say

A parser reads each sentence. Action verbs with an on-screen label become clicks or typing; “here you'll see” becomes a spotlight; everything else is narration.

02

Matching by label and role

Visible text, aria-label, placeholder, title, and input type, with synonyms, fuzzy matching, and a preference for whatever dialog is open.

03

Pins that survive redesigns

data-testid, ids, aria-labels, and stable class names, in that order, so a moved button is still found.

04

Readiness, not timers

It waits until a control is enabled, not busy, and still, instead of sleeping for a guessed number of seconds.

05

Timing from the voice

The click lands on the trigger word in the sentence. Word timestamps from the narrator decide when.

06

A second opinion, only in rehearsal

Chrome's on-device model, or a cloud model if you switch it on, settles look-alike picks and reviews drawings before they happen. Playback never waits on a model.

For developers

Your app can make itself easy to walk.

The app guide is a one-page checklist: names on every control, unique labels, dialog roles, data-testid on the things a tutorial uses. Paste it into your AI assistant's instructions and every new screen arrives walkable.

Walkthrough in three steps

1

Describe the walk

Chapters for each area, one sentence per step.

2

Rehearse

Pin every step. Point at anything the words did not catch.

3

Run and record

Play to check. Export to record. Run again any time.

Product walkthrough videos: the questions people ask

Is there a tool that generates walkthrough videos automatically?
Yes, this one. The automation is in the performance: given a script, Tuti Tuti finds each element on the live page and performs the step while the sentence is spoken. It is not generating fake footage; it is driving the real app.
How is this different from Scribe or Guidde?
Those capture what you do and turn it into steps or a video; you still do it. Tuti Tuti starts from the words and does the steps itself, which is what makes the result repeatable after the app changes.
What if my app's buttons have no labels?
Rehearsal tells you exactly which sentence missed, and you can point at the element to pin it. The lasting fix is to give the control a name, and the rework file tells your AI which controls and which names.
Does it handle dialogs, menus, and drawers?
Yes. It prefers whatever dialog is open, discounts the page behind it, scrolls containers to reach items, and waits for drawers to finish sliding before it clicks.
Can it type into fields and pick from dropdowns?
Typing, yes, one character at a time with real input events so React and Vue forms update. Dropdowns are clicked like anything else; a select sentence names the option.
Does it use AI to drive the app?
Rules decide what to do; they are deterministic and tested. Models are optional second opinions during rehearsal only, so a walkthrough plays the same way every time and never waits on an API.

Describe the walk. Watch it taken.

Install, paste, rehearse, export.

Add to ChromeCreate an account

Tuti Tuti - powered by Pandacat.