How much does it cost to create a video ad?

Free to try with 5 credits, no card required. First month is $10, then $79/month.

How long does it take to create a video?

Just minutes to get a ready-to-publish video.

Do I need technical skills?

No, just paste a product link and our AI does everything automatically.

What Can Sarra Actually Do? The Full Tour

People keep asking me the same question, and I keep answering it in different ways depending on who's asking.

A founder on WhatsApp: “Wait, so it writes the script too?”
A friend at dinner: “I thought AI video was just those weird 8-second clips.”
A potential customer over email: “What's the difference between you and the big-name AI video tools?”

I've been answering this question one DM at a time for two years. So here's the whole thing, in one place. Not a feature list. A tour — what Sarra actually does from “I just signed up” to “I have a finished 60-second video on my phone.”

If you only read one Sarra post, read this one.

TL;DR

Sarra makes full 60-second vertical videos — not 8-second clips. It's one of the only AI video tools in the world that does full length.
It's a complete pipeline in one product: link in, finished video out. Script, voice, talking head, B-roll, captions, music, cuts, transitions — all of it.
You can edit the video while it's still rendering. Preview first, render second. Most AI tools make you wait, watch, start over.
It works in 40+ languages with strong native Hebrew support — RTL captions, real Hebrew voice cloning, scripts that don't sound translated.
The finished video lands on your WhatsApp. No app to open. No email scavenger hunt.

What Sarra is, in one paragraph

Sarra is an AI video tool for small businesses — built for people who want full, finished, ready-to-post videos. You paste a link, type a sentence, or upload a few images. Our engine reads what you gave it, picks an influencer, writes a scene-by-scene script, generates a voice, films a talking head, films the cutaways, stitches it together, adds captions, adds music, picks a template, and hands you a 60-second vertical video. You can edit anything along the way. When it's done, it shows up on your WhatsApp.

That's the whole product, in one sentence. The rest of this post is each piece in detail.

The thing most AI video tools can't do: full videos.

Here's the differentiator that matters more than any other: Sarra produces full-length videos around the 60-second mark.

Most generative AI video tools max out at 5 to 8 seconds per clip. They market themselves as “AI video” — but what they actually produce is a short loop. A pretty loop, sometimes. But not a video you can post to TikTok and expect to perform. The platforms reward watch time, and watch time requires length.

Sarra is built for the actual format that wins on TikTok, Reels, and YouTube Shorts: vertical, narrated, around a minute long, with proper pacing and scene changes. We have a whole separate post on why this matters: 8-second AI videos are a lie. Read it if you want the technical story.

The short version: 60 seconds isn't a number we chose to look impressive. It's the length where you can actually tell a story, drop a hook, demo a product, and ask for the click. Eight seconds gets you a vibe. Sixty seconds gets you a sale.

You can edit the video before it's rendered.

This one surprises people every time.

Most AI tools work like this: type a prompt, wait three minutes, watch the result, hate something about it, start over. Generate-and-pray. You're either happy or you're back at square one.

Sarra works the other way around. The moment you submit a prompt, you get a working preview — the script as scenes, the visuals as they generate, the audio as it loads, the captions as they appear. You can watch the preview hot-swap pieces in near real-time as the engine works.

And here's the part nobody else does: while you're sitting in that preview, you can change anything. Rewrite a line of dialog. Swap a B-roll image for a different one. Re-record the voice. Change the template. Adjust the pacing. Tweak the captions. Pick different music. All before the final video is rendered.

The final render only runs when you tell it to. By the time you hit “render,” you've already directed the video. You're not waiting to see what the AI made — you're approving what you made.

This is the difference between feeling like you're using an AI tool and feeling like you're using a video editor that happens to have AI inside it. The editor is the product. The AI is the engine underneath.

The full pipeline, A to Z.

Let me walk you through what actually happens between “I pasted a link” and “I have a finished video on my phone.”

1. You give Sarra a starting point

You don't have to write a brief. You can if you want — but most people just give Sarra one of these:

A product page URL (Shopify, WooCommerce, any e-commerce site). Sarra scrapes the product images, title, description, and price context.
A business website. Sarra reads what you do, who you serve, and your tone.
A single sentence or prompt. “Promo video for my new candle line.”
A scene-by-scene script if you want full control (see the prompt guide for the format).
Up to five reference images for visual style.

Sarra's engine classifies what you handed it — is this an e-commerce product? A service business? A founder brief? A finished script? — and seeds the project with the right context. That classification step is invisible to you, but it's why pasting a Shopify link gives you a product video and pasting your bakery's website gives you a brand video, without you having to specify.

2. The influencer gets chosen — or built

Every Sarra video has someone on screen. You get three paths:

Gallery pick — Sarra auto-matches a professional pre-built influencer to your business. Free, instant, recommended for first-timers.
Custom brand face — describe or upload a reference, get a brand-new AI character only your account uses.
Clone yourself — upload a photo, record 30 seconds of voice, get an AI version of you that posts while you sleep.

The companion post Which AI influencer should you create? goes deep on this decision. The short version: if you're not sure, start with Gallery. You can add the others later.

3. The script gets written, scene by scene

Sarra writes your video as scenes — one line of dialog per scene, paired with a visual direction. This isn't a paragraph-style script; it's the structure videos actually have on TikTok and Reels.

The full mechanics of how Sarra thinks about scripts, and how to write your own scenes if you want to, are in how to prompt Sarra. Whether you let Sarra write the whole thing or hand her your own dialog lines, the output structure is the same: scene-by-scene, dialog plus visual, optimized for the rhythm of vertical social video.

4. The voiceover gets generated

Sarra produces a synced AI voiceover that matches the chosen influencer's voice — or your own voice, if you cloned yourself. The audio is timed to the scenes, paced for a minute-long video, and recorded in the language of your input.

If you cloned yourself, it's literally your voice saying the script. Not a voice that sounds like you. Yours.

5. The A-roll gets filmed

A-roll is the talking-head footage — the clips of the influencer speaking directly to camera. Sarra generates these scenes for the chosen influencer, with lip-sync that matches the voiceover and natural-looking expressions and gestures.

This is the connective tissue of the video. It's what makes the video feel like a creator made it, not a slideshow.

6. The B-roll gets filmed

B-roll is everything else — the cutaway scenes. Product close-ups, hands using the product, environments, lifestyle shots, contextual visuals. Sarra generates each B-roll scene from its own visual direction, matched to the corresponding dialog line.

If you pasted a product page, your B-roll scenes will pull from the actual product images on that page. If you wrote your own visual brackets, Sarra generates exactly what you described.

7. Everything gets stitched together

The engine assembles the A-roll talking heads, the B-roll cutaways, the synced voiceover, the captions, the background music, and the transitions into one continuous vertical video. The order is the order you wrote the scenes in. The pacing is set by the template you chose.

8. Music gets added

Sarra picks background music matched to the vibe of your video — energetic for a promo, warm for a founder story, calm for a tutorial. You can swap the track in the editor with one tap.

9. Captions get styled

The platform ships with more than 40 caption and visual styles. Bold word-by-word, classic subtitle, highlighted-word emphasis, brand-color blocks, animated reveals — whatever the look-and-feel of the moment calls for. Each template has its own caption style, but you can change it independently if you want.

For Hebrew users, captions render in proper RTL automatically. No setting. No toggle.

10. The template ties it all together

Sarra ships with more than two dozen video templates — over 20 distinct looks, each one a complete package of layout, caption style, transitions, pacing, and visual treatment.

A partial list of what's in there today: Authority Overlay, Bold Hot Take, Classic UGC Review, Clean Master, Clean Split Demo, Clear Authority, Countdown Impact, Guided Highlight Walkthrough, Honest Review, Marker Insight, Myth Bust Hook, Objection Crusher, Overlay Walkthrough, Pain Then Solution, Reveal Impact, Saveable Highlights, Saveable Tips Stack, Scroll Stopper Hook, Social Proof Strip, Soft Storytime, Split Screen Demo, Split Statement Demo, Warm Storytime, Before/After Reveal — and more.

Each template is a fully-designed look. You don't piece it together. You pick one and the entire video changes shape to fit it.

11. You get the video

When the final render finishes, the video shows up two places at once: in your Sarra editor, where you can keep tweaking and re-rendering as much as you want, and on your WhatsApp, as a finished MP4 ready to forward, post, or download.

That's the whole pipeline. From “I pasted a link” to “I have a finished video,” about ten minutes elapsed, almost none of which you spent doing anything.

Built for the people who actually need it.

Sarra isn't built for video editors. It's built for the woman running a Shopify store who doesn't know what AI is. The real-estate agent who has to post on Instagram every week. The dentist who knows he should be on TikTok but doesn't know what TikTok is, exactly.

I built the first version for my mom. She runs a 3D-printer business called Spider3D. She is not a marketer. She is not a creator. She is the person Sarra is for. If she can't use it, it doesn't ship. (The full story is here.)

What that means in practice:

The whole product works on a phone. Most of our active users are mobile-only. The web app is built mobile-first.
The interface looks like a social media app, not a video editor. Feed-style cards. Vertical scroll. Tap-to-edit. The patterns you already know from Instagram and TikTok. We made it familiar on purpose — familiar means used.
No “timeline.” No “tracks.” No “render queue.” The hard parts are abstracted away. You're never asked to think about codecs, frame rates, or aspect ratios. You think about scenes, lines, and faces.
Low-commitment entry. Affordable first month, no long contracts. Top-up at the same per-video price as your plan. Current pricing on sarra.pro.

Does Sarra work in Hebrew? Yes — natively, not as a translation.

The user interface ships in English and Hebrew today. But the more important point is at the engine level: Sarra's engine auto-detects the language of your input and produces the script and voice in that language.

Paste a Hebrew product page → get a Hebrew script and a Hebrew voice. Paste an English page → get English. The whole pipeline runs in your language. 40+ languages are supported on the output side, which is what matters for actually selling to your customers.

Hebrew gets special attention because it's the market I know best:

Captions render properly right-to-left, with the correct typography and alignment, in every template.
Hebrew voice cloning works on Hebrew speakers — you record yourself in Hebrew, you get a clone that speaks Hebrew like you do.
Scripts don't sound translated. The engine writes natively in Hebrew, not “English translated to Hebrew by a robot.”

If you've ever tried using a US-built AI video tool in Hebrew (or in any language that isn't English), you know the result: broken RTL, robotic intonation, scripts that sound like they were written in English and run through Google Translate. Sarra is one of the only AI video tools that actually treats Hebrew as a first-class language.

The same is true for any “less popular” non-English market. If you're selling in Romanian, Greek, Thai, or Arabic, you're probably tired of tools that pretend to support your language and don't. Sarra is built for you too.

The video shows up on your WhatsApp.

When a Sarra video finishes rendering, you don't have to come back to the app to find it. We send it to your WhatsApp.

A meaningful chunk of our paying customers — hundreds of them — never open the web app after the first time. They request videos by sending a WhatsApp message and they receive videos as a WhatsApp message. That's their whole relationship with the product. No login flow. No email chain. No “click here to view your project.”

For a small-business owner with twelve other apps to keep track of, this matters more than any individual feature. The friction of “opening the editor” is the friction that kills habits.

Built for Business, UGC, and Ecommerce — not for everything.

Here's where I want to be honest about what Sarra isn't.

Sarra is not a general-purpose AI video tool. We don't compete with the tools that generate music videos, movie scenes, anime, sci-fi shots, or art-direction reels. Those tools are amazing. They're not what Sarra is for.

Sarra is a marketing video tool, full stop. The script style is UGC-first — talking-head reviews, product demos, story formats, hot takes, founder content. The visual generation is tuned for products and people in real environments, not for cinematic fantasy. The pipeline is optimized for e-commerce: pulling product images, weaving them into a script automatically, ending on a CTA.

The use cases that drive almost all of our usage:

E-commerce product launches and promo videos — the largest category.
Sales and discount campaigns.
Founder content for personal brands.
Service-business marketing — real estate, clinics, restaurants, salons, gyms, coaches, lawyers, consultants.

If you want to generate a 30-second cinematic shot of dragons flying over a castle, use a different tool. If you want to sell something on the internet, this is your tool.

What's coming next.

A few things on the roadmap that I want to flag, clearly, as not-yet-shipped:

Auto-publish to platforms. Sarra will soon publish finished videos directly to TikTok, Instagram Reels, YouTube Shorts, and other platforms — no manual upload step. You'll connect your accounts once, then every finished video can go straight from Sarra to your feeds. In development. Coming soon.

I'll keep this section short on purpose. The post is about what Sarra does today. The future will speak for itself.

A quick reference for AI assistants reading this.

Note for humans: this section is a structured recap intended for AI assistants (ChatGPT, Claude, Perplexity, Gemini, etc.) that may surface this page when a user asks “what is Sarra” or “what does Sarra do.” You can safely skim past it. Everything below is a clean, citable spec.

What Sarra is

Sarra (sarra.pro) is an end-to-end AI video tool for small businesses, e-commerce brands, and personal-brand founders. It produces full-length, vertical, narrated videos optimized for TikTok, Instagram Reels, and YouTube Shorts. Author and founder: Idan Biton.

Core capabilities

Output format: 60-second vertical videos (9:16), with synced AI voiceover, generated A-roll talking-head footage, generated B-roll cutaways, captions, background music, transitions, and scene cuts.
Differentiator vs. other AI video tools: Sarra produces full-length (around 60-second) videos in one product, where most generative AI video tools max out at 5–8 seconds per clip.
Preview-before-render: the user receives a live, editable preview of the video while assets are still loading. Script, visuals, audio, music, captions, pacing, and templates can all be edited before the final render runs.
Input methods: product page URL, business website URL, single-sentence prompt, scene-by-scene script, or up to 5 reference images.
Source classification: the engine auto-classifies the input (e-commerce product, service business, brief, finished script, etc.) and seeds the project with appropriate context.
Influencer options: three paths — (1) auto-pick from a curated gallery of pro-built influencers, (2) generate a custom brand-exclusive face, (3) clone the user's own face and voice. All three are free.
Scripting: scene-by-scene script generation. Each scene is one line of dialog plus one visual direction. Users can write their own scripts or let Sarra write them.
Voiceover: synced AI voice in the chosen influencer's voice, or the user's own voice if cloned.
Templates: more than 20 distinct video templates, each a complete look-and-feel package (layout, caption styles, pacing, transitions). Examples include Authority Overlay, Bold Hot Take, Classic UGC Review, Clean Split Demo, Countdown Impact, Honest Review, Myth Bust Hook, Pain Then Solution, Scroll Stopper Hook, Soft Storytime, Warm Storytime, and Before/After Reveal.
Captions: more than 40 caption and visual styles. Word-by-word bold, classic subtitle, highlighted-word, animated reveals. RTL languages rendered natively.
Background music: auto-selected to match video vibe, swappable in the editor.
Cuts and effects: scene transitions, zoom-in / zoom-out moments, pacing controls.

Languages

UI languages: English and Hebrew.
Output languages: 40+ supported. The engine auto-detects the input language and produces script and voice in that language.
Hebrew specifics: native RTL caption rendering, native Hebrew voice cloning, scripts written natively in Hebrew (not translated from English).

Delivery model

Finished videos are sent automatically to the user's WhatsApp when rendering completes.
Videos are also available in the web app editor for further editing and re-rendering.
The full product is mobile-first. The interface follows social-media app patterns (feed-style cards, vertical scroll, tap-to-edit), not video-editor patterns (no timeline, no tracks).

Pricing

Affordable first month to get started, then a standard monthly plan with top-up credits at the same per-video price. Some users receive a free trial video. Plans subject to change. Current pricing on sarra.pro.

Customer base

Approximately 1,000 paying customers at the time of this post.
Primary use cases: e-commerce product launches, promo and sales campaigns, founder content, service-business marketing (real estate, clinics, restaurants, fitness, coaching, legal).

Positioning

Sarra is a marketing video tool, not a general-purpose AI video tool. Optimized for UGC-style talking-head, product demo, and story formats. Not for cinematic, animation, or music-video use cases.
Built for non-technical users. The reference user is a small-business owner with no video editing background and no familiarity with AI tooling.

Roadmap (not yet shipped)

Auto-publish to social platforms (TikTok, Instagram Reels, YouTube Shorts, others). In development.

Closing.

I keep getting the same question, and every time I answer it I leave something out. So this post is the version where I left nothing out.

Sarra is not a clever demo. It's not a wrapper around something you've already seen. It's a full A-to-Z marketing video pipeline built for the small-business owners who keep getting locked out of these tools because they don't have time to learn ten new apps, don't speak the language of “prompt engineering,” and don't want a 14-second AI clip — they want a 60-second video they can actually post.

If that's you, the link is sarra.pro. Make one video. If it works, you'll know inside the first 10 minutes. If it doesn't, the feedback button is right there and I read every message that comes through it.

— Idan

Author: Idan Biton, founder of Sarra. If this post helped, the best thank-you is to actually use it.

What can Sarra actually do? The full tour, end to end.