Skip to content
New · the open voice benchmark is liveRead it
cantari
For SaaS and product marketers

Walkthrough VO that keeps pace with the release train.

The product changes faster than a voice session can be booked. Generate the walkthrough read with real style and speed controls, re-render when the UI does, and lay a generated bed underneath, all before the feature flag flips.

No credit card · Real engines · The audio is yours

Painted transcription desk with a typewriter and tape recorder
Generated voice
MP3 + WAV · yours to export
Worked example

Feature intro: twenty seconds over a screen recording

Script fragmentMAI Voice 2

This is Boards, the newest part of your workspace.

Drag any report onto a board and it stays live. The numbers update themselves while you present.

Share it with one link. Viewers do not need an account. Let's build one from scratch.

Line 2, real MAI Voice 2 output, unedited.

MAI Voice 2, voice Harper: real style and speed controls, so the same script renders cheerful for the launch cut and measured for the documentation version.

The honest arithmetic · about 1,000 characters is a minute of speech
~330
characters in a 20-second feature intro
0.5-2x
real speed range on MAI Voice 2
4
styles beyond neutral: cheerful, excited, sad, angry
Why this is hard

What Product Demos actually needs.

We would rather name the friction plainly than pretend it away. Here is the problem this page is about.
The honest problem

Demo VO has the shortest shelf life in marketing: the UI moves, the button gets renamed, and the polished video is suddenly showing the old flow. Re-booking a narrator for every release means the demo library is always a version behind, so teams ship silent screen recordings instead. The fix is a read you can regenerate the day the release lands, with controls for energy and pace so the same script fits a launch cut and a docs walkthrough.

Sound familiar?

Sales wants a walkthrough video for the new feature by Friday, and the founder's voice memo is not going to cut it. The product changed twice while the last demo was being voiced. You need VO that keeps pace with the release train, with enough energy to carry a screen recording.

How Cantari helps

Real features, mapped to the job.

Every item here works today, or says plainly where it is still in progress.

Real style and speed controls

MAI Voice 2 accepts genuine style and speed parameters, with an intensity dial, so the launch cut can run brighter and the tutorial can slow down, from one script.

Keep pace with releases

When the UI changes, re-render the affected lines in minutes. The demo library stops trailing the product by a version.

Beds for the background

Sound & Music composes instrumental beds live today, so the walkthrough gets an understated track without a stock-library hunt.

Own the VO

Export MP3 or WAV with commercial rights and no watermark, cleared for the site, the deck, and the campaign.

The workflow

How it goes, step by step.

Step 1: Script to the screen recording

Write the VO against the captured flow, one line per beat, so the read lands on the right click.

Step 2: Set style and speed on MAI

Pick Harper, set the style and pace for this cut, and generate. The controls are real engine parameters, not a filter.

Step 3: Re-render what the release broke

When a button gets renamed, regenerate that line alone and drop it back on the timeline.

Step 4: Add a bed and export

Generate an instrumental bed in Sound & Music if the cut needs one, then export everything. It is yours.

Library notes

Keeping a demo library honest across releases.

Version the script with the feature

Store each demo script next to the feature's release notes and stamp both with the version. When support links a walkthrough, anyone can see which build it shows, and when the UI moves you know exactly which lines to re-render instead of re-watching every video to find out.

One script, two energies

The launch cut and the docs walkthrough should not sound the same even when the words match. Render the demo VO twice on the styled engine, brighter for the campaign and measured for the help center, and each surface gets the read its audience actually expects.

Recommended engine

Start with MAI Voice 2.

MAI Voice 2 has real style and speed controls with an intensity dial, which is what demo VO needs: the same script at launch energy or tutorial calm, set by you rather than luck.

MAI Voice 2Styled voice - real controls

Microsoft voice with real style and speed controls.

Quality Elo
1007 *
Latency
2426 ms (measured 2026-06-10)
Languages
1
Rights
Commercial use; outputs are yours
Style controlsEnglish
Hear a line for this use case

Meet the new dashboard. Everything updates live, and yes, you can share it with one link.

Real MAI Voice 2 output, recorded unedited.

The honest answers.

What Cantari can and cannot do for product demos today, in plain language.

What does the same script sound like in different styles?
MAI Voice 2 accepts cheerful, excited, sad, and angry styles plus a neutral default, with an intensity control and a 0.5x to 2x speed range. Those are real engine parameters we verified, so two renders of one script can genuinely differ.
Can I add background music to the demo?
Yes. Sound & Music is live today and composes instrumental beds and atmospheres from a plain-language description. It does not do one-shot effects like a single click sound yet, and we say that rather than pretend.
Is the VO cleared for the website and ads?
Yes. Everything you generate is yours commercially, worldwide, with no watermark and no attribution required.
Keep exploring

Try Cantari for product demos.

Free to start, no credit meter. Open the studio and hear it for yourself.