One ad line, a 1-minute quiz, and a 5%-off annual plan: Elva sold "hundreds of unedited clips on your phone" into a subscription earning about $90k a month

Elva is an app you edit video with by talking: pick clips from your camera roll, say what feel you want, and it selects the shots, cuts to the beat, adds music, and captions. Users are unhappy with the results, and most recent iOS reviews are one or two stars. Still, a third-party estimate puts its iOS monthly revenue at about $9k in 90, up from just a few hundred dollars half a year earlier. The product didn't do much of the selling; a paid acquisition funnel did: Meta video ads, a web quiz, a discounted annual plan, aimed at women aged 35 to 54 who don't edit. The playbook is worth studying, but the product itself isn't worth chasing.

A UK user who paid after seeing the ad

In 2026 June 19, someone gave Elva one star on the UK App Store with a single word as the title: "Scam."

"I saw the ad, got excited, and bought this app. It doesn't work. I tried many times and it still doesn't work, and I couldn't find anyone to help. They took the money and gave no service. No wonder there's no free trial."

Sep 9 another one: "Just paid £18, no free trial... no matter how many times I try, it doesn't cut the way I tell it to." On Oct 10, a user in the US store wrote that she emailed the support address to cancel her subscription and get a refund, and neither worked. She also couldn't find a cancel option anywhere in the app.

These reviews describe the same path: you see the ad, you pay, and only then do you discover the product doesn't work well. Elva's revenue depends on the first half of that path.

What it is

The first line of the store page: "Stop spending hours editing." You pick a batch of videos and photos from your camera roll and say "turn last weekend into a travel reel." It analyzes the footage, picks the highlights, arranges the structure, adds beat-synced music and transitions, and generates captions from the voice, with optional AI narration. If you don't like it, keep talking: "Make it punchier," "change the music." There's also a capture mode that gives real-time framing and lighting tips while you shoot.

The developer is Elvalabs LTD, which LinkedIn lists as a Cyprus-based company with 5 people. The iOS version launched in 2026 March 3, with 392 ratings on the US store, averaging 4.09 stars (App Store, 11-06). The Android version launched around September (the store's listed date is Jul 28, probably edited later). By Nov 6 it had 5,873 installs, with 8 more in the last 972 days, about a hundred a day. The brand, motion design, store images, and paywall were all done by the design agency Lazarev, which won a People's Voice Webby Award in 2026 for AI visual design.

Where users come from: one line, 50 ads

Meta video ads are the main channel. A third-party ad monitor (Airbridge, published 10-19) found Elva's Facebook page running 50 ads in October, versus a median of 21 for photo and video peers. From Oct 7 to 17, active ads went from 35 to 58. The top 13 creatives are all video, and 11 of them use the same copy: "Stop wasting hours editing... just say what you want... No skills. No timeline. Just results." The estimated audience is 69.5% US and 16.7% Spain; 61% women, 35–44 year-olds 37%, 45–54 year-olds 31%.

These aren't content creators. They're ordinary people with phones full of kids, trips, and parties they've never edited. Most have never heard of CapCut or Captions, and they don't mind a price of over $100 a year. Some US reviews are in Spanish, and some users edit "getting ready to make 2 lb pork rice" videos, which shows who the users are.

The ads don't send people straight to the store. They land on a quiz page first. The landing page is quiz.elvalabs.ai, with the headline "Turn your clips into cinematic videos in minutes" and the line below it "1 minute quiz — we'll set up Elva just for you". The page images load from FunnelFox's CDN, a web2app funnel tool. Lazarev's case study spells out the quiz logic: each question looks like it's collecting preferences, but really it makes you feel what it costs not to act. The original line: "How many unused videos do you have on your phone right now? Hundreds? That's a lot of memories nobody's ever seen." By the time the quiz ends, people already feel they owe themselves something, and the paywall comes next.

The paywall anchors on a high-priced annual plan. The iOS in-app purchases run from $9.99 to $149.99 across 10 tiers, many labeled "50% OFF Annual" or "50% OFF First Month." The $129.99 and $149.99 annual plans make the discounted price look like a deal, and there's no free trial. Android prices run $1.99–$129.99. Payments made on the web don't go through the store, so outside revenue estimates are probably low (inference).

They have a dedicated Head of Growth. The person who joined in 2025 December was previously a monetization consultant focused on onboarding, paywalls, pricing and web2app work, wrote "How to hack your Web2App funnel", and spoke at the 2026 Business of Apps conference in Berlin. After she joined, revenue went from $598 in March to a September peak of about $90k (Airbridge estimate). The timing lines up, but that doesn't mean she's the reason for it.

Almost no organic traffic. The official TikTok account has 2 followers. X has no discussion in the past half year, and Reddit and Product Hunt have no launch posts. The Instagram account @elvalabs.ai has 9,057 followers, runs "Comment VOICE" giveaways for links, and has one paid creator partnership, small in scale.

Reputation: 4.75 stars on Android, while iOS reviews are almost all one-star

Of 570 Android ratings, 514 are five-star, averaging 4.75 stars. But in the last 8 days there were 103 new ratings, and only 1 of them included text. About 10% of new installs left a rating, a very high share, most likely from an in-app prompt asking for a review (inference).

User-written reviews on iOS are a different story. Airbridge sampled 118 recent reviews, averaging 2.47 stars. The recurring complaints fall into these categories:

"I kept asking it to remove the music, it said okay, and then it just didn't." — US store, 10-14 "Too slow, I have to render over and over… regret buying it." — US store, 10-31 "Takes longer than editing by hand in CapCut. I asked it to delete one clip three times, and it deleted the one I said not to." — UK store, 8-27

Five-star reviews too: "I could never edit anything before. Now it looks like a pro cut it in minutes." (US store, 11-01) The people praising it are almost all folks who've never edited a video, which is exactly the crowd the ads are chasing.

So the money comes from first-purchase conversion, not retention. Renewal and refund rates aren't visible from outside. On the Google Play video top grossing chart, it fell from #93 in Nov 1 to #118 in Nov 6. Estimated iOS revenue peaked in September, so it may already be past its peak (inference).

Who it's competing with

Only does talking-head and UGC cleanup, narrower use case

"Say one sentence, get a photo reel" is not a scarce idea. Reelful and Elva launched the same month, and by 8 and October Narrative, Ody, CutAI and Picsart had joined. On the other side, CapCut, Edits, Premiere and the system photo apps are all free, and they're all adding AI. Elva's edge is mainly who it sells to, how it's packaged and how it charges. The editing itself isn't better than the others, and user reviews make that clear.

(US App Store rating and Google Play installs, early November.)

If it were you

One-line verdict: The thing worth copying is the funnel: a video ad with a hard promise, a web quiz that amplifies the cost of not acting, and a discounted annual plan. The category itself isn't worth going head-to-head in. It's already crowded, free big-tech tools sit underneath it, and AI editing is the hardest part to deliver. Most of Elva's bad reviews come from exactly that.

Who it suits: small teams that run Meta ads, can build a web2app funnel, and can work out how long it takes a first purchase to pay back; anyone who has an AI use case where regular people have the footage but can't make the result themselves.

One-line verdict: The part worth learning is the funnel: a video ad with a strong one-sentence promise, a web quiz that amplifies "the cost of not acting", and a discounted annual plan. The category itself isn't worth attacking head-on. It's already crowded, free big-company tools sit underneath, and AI editing is the hardest part to deliver reliably. Elva's bad reviews almost all come from there.

How to do it:

• Narrow the funnel to scenes where AI can deliver a good result, instead of building another general-purpose AI video editor. Examples: pet memorial videos, yearly highlight reels for kids, wedding and birthday photo slideshows, small-shop promo videos. The output format is fixed, so templates give stable, good-looking results. You don't need AI to understand every instruction.
• Do the opposite of what it gets roasted for. Its bad reviews cluster around three things: ignoring instructions, slow renders that crash, and hard cancellations and refunds. Offer a manual edit fallback, make the first video free to finish before charging, and put a prominent cancel button in the app. That way your paid acquisition money doesn't get eaten by refunds and bad reviews.
• Take the funnel into a narrower niche where AI can deliver well, instead of building another general AI editor. Examples: pet memorial videos, yearly compilations of a child's growth, photo slideshows for weddings and birthdays, small-shop promo videos. The output format is fixed, so templates can look consistently good without relying on AI to understand every sentence.

How to get users: Copy its skeleton. For Meta video ads, use one promise only, like "Stop spending hours editing, just say it," and keep testing new visuals around it. On the landing page, run a 1-minute quiz, with questions like "How many unedited videos are on your phone?" Target women over 35 and parents, and stay away from the young creator crowd that CapCut reaches. A web2app tool (FunnelFox-style) works out of the box, and web checkout also saves you the store commission.

When the window closes: The funnel playbook won't expire, but the "one-line AI video" pitch is becoming a free feature at big companies. Once the system photo gallery and CapCut make conversational editing a default, the room to charge gets squeezed. If you want in, get your materials and payback model working before the first half of 2027 (inference).

What it looks like when you've misread it: you've tested creative for a month or two and first orders still don't cover acquisition cost; or refunds and one-star reviews are piling up faster than revenue; or in your chosen use case, the system photo app or CapCut already does the same thing for free. If any of these shows up, stop.