Get paid to film yourself doing chores: Reps turns "recording training videos for robots" into a money-making app

Reps is a US-store iPhone app from Born Labs in New York. Users film themselves sweeping floors, peeling vegetables, and wiping tables from a first-person view, and also snap receipts, record voice, and share location. Each task earns money into the in-app wallet, and the footage is sold to AI and robotics companies. It launched in 2026 Jul 20 and picked up about 15k ratings in roughly three and a half months. Its rivals have raised tens of millions of dollars in funding yet have only a few dozen ratings on the App Store for the same idea. Reps didn't win on tech. It won by turning gig work into an app you install and start earning from, then using a "subscribe before you can cash out" model to recover acquisition spend from users. But in early November it pulled the most valuable task category.

A review asking "where did my Strap tasks go"

In Nov 2026 1, a US user left Reps a 5-star review on the App Store, titled "I think I like it but…":

So I easily got up to $30.00, bought a strap on my own, got it approved, did a couple of successful "strap" videos. The app recently updated and now there is no mention of the strap or means to earn making strap videos anywhere on the app (which are the more lucrative quests). What happened here?

Several more the same day: "They took away strap tasks & the app keeps glitching" (2 stars); "I've been filming chores with a head strap, and after an update all my earlier strap videos vanished, and even the option is gone. Only quest is left" (10-31,1 stars).

"Strap" is Reps' most valuable task: you fix your phone to your forehead, film chores from first-person view, and get paid by the minute. It is also the centre of Reps' whole payment design. The three things in this review, reaching $30, buying your own strap, and strap tasks disappearing after an update, sum up how the app took off and where it went wrong.

What it is

The store description is blunt: "you get paid to record yourself completing simple, everyday tasks — vacuuming a floor, sanding a wooden surface, peeling vegetables," and your recordings "help teach the next generation of AI and robotics."

Based on the terms and the official site (earnreps.com), the app has two kinds of task:

• Quests (doable on your phone): photograph receipts and bills, photograph the buttons on your fridge and microwave, record a voice note, do a 3D room scan, answer video interview questions. They take a few minutes after install and pay from a few cents to a few dollars each.
• Strap tasks: film chores wearing a head strap, paid by the approved minute. Chatter in side-hustle groups puts the hourly rate at $15–21, with a $20 bonus for completing the first hour (from a Facebook group post).

There is also a "location sharing" program: set location permission to "Always," and you get a small amount of money each day. The terms state that receipt, location and interview data will be turned into "aggregated consumer and in-store insights" for brands. So it isn't only selling robot training data. It is also doing a consumer-research kind of data business.

iPhone only, US only. Not on Google Play.

How big is it really?

Compare it with the rivals and the picture is clear. Shift (MicroAGI), which does the same thing and raised the most money, took a $55M seed round in July 2026 and claims over $100M in half-year ARR. Yet its iOS app has only 40 ratings, averaging 2.65. Luel, which raised $31.2M, has just 9 ratings on the US iOS store (all App Store, 11-06). By App Store scale, Reps is first in this category.

But don't overrate it. It has never charted on Top Grossing, which suggests it doesn't even reach the top 200 of the US Utilities grossing chart, so subscription revenue can't be large (inference).

Why now

In 2026, robotics companies badly needed first-person footage of people working in real homes, and they were willing to pay by the hour. XDOF raised $70M, MicroAGI $55M, Luel $31.2M. Wired published "I Spent a Week Recording Myself Doing Chores for Money" in May 26, and CBS covered Shift. "Film chores to earn money" went from a lab's procurement need to something ordinary people search for.

The money sat with the buyers, the curiosity sat with ordinary people, and what was missing in between was an entry point ordinary people could start using immediately.

What it got right: turning gig work into an earn-money app

Rivals all built it as "hiring gig workers": Shift made you buy a $34 strap first and only refunded it after 10 hours of work, and required slow motion with no hands in frame. HighFive only let you cash out once you hit $100 a week, and refunded the strap only after 15 hours. Contracted AI and micro1 required interviews and contracts.

Reps built itself on the model of Freecash-style "earn pocket money on your phone" apps: no interview, no deposit, quests ready the moment you install, a wallet balance that climbs bit by bit, plus streaks, multipliers and membership cards. Even the name shares its wording with Freecash's "Get Paid Real Money", and it rides the "get paid" search terms (inference).

Then comes the payment step, the most critical part of the whole design:

1. Quests quickly bring you to $20–50;
2. To cash out for the first time, you have to accumulate 60 minutes of approved Strap footage. Quest and referral earnings don't count;
3. Accounts created after Sep 10 must subscribe to Ultra at $9.99 / month before they can claim the free strap.

Let you earn first, then charge you to unlock. The data buyers haven't paid up yet, but users' subscriptions have already recovered the acquisition cost.

The cost is also written in the reviews: "SCAM!!!! If you're ready to earn and then have to pay 9.99 to unlock the strap to get your money" (10-27); "Makes you pay $10 to earn your money. Doesn't say this upfront" (10-26); "I bought the $9.99 strap & subscription LAST MONTH…. they no longer allow me to use one I bought on Amazon. So I have $50 I can't even cash out" (10-25). With this many negative reviews, the average still sits at 4.86. Many of the most recent 5 stars say "Just downloaded", "Can't wait to start", "Haven't tried yet", which suggests the rating prompt fires during onboarding (inference).

Where users come from

1. Paid UGC accounts on TikTok. Search "reps app" and most results follow the same template: "#Reps #fyp #sidehustle #easyhustle #app". @isabborgees posted 372 times with a bio that reads only "Go download Reps‼️" plus an App Store link. @blair.ugc (24.6k followers) describes herself as "UGC Creator… Paid Collabs" and also carries Reps. @nicholassc01 posted an Aug 25 video with 99.5k views. All the links point straight to the App Store, not to personal referral codes, which suggests brands are paying for this content.
2. Facebook side-hustle groups and "AI training" groups. Groups like "The AI training creators community in USA" post rankings and tutorials that put Reps, Shift and HighFive side by side. Reps is sold as "VERY easy", and you can "start earning quests before your strap arrives."
3. Two-tier referral commission. Early September there was none (the Sep 9 ranking post says "REPS DOESN'T HAVE A REFERRAL LINK"). It was added later: you earn a bonus when your referral completes their first job, plus a cut of their referrals' earnings and their referrals' referrals. By October people were posting their codes in groups and bragging "someone used my code and I got $20."
4. Search demand from the trend. Wired and CBS coverage made "get paid to do chores" a search term, and Reps caught it. The website's download button uses an AppsFlyer attribution link, so there's a paid-acquisition setup behind it, but I found no record of them running ads (unverified).

It's the usual reward-app playbook, moved to a new theme.

Who it's competing with

Plenty of companies do this, and the money is on the buyer and gig-platform side. Only Reps and Amelia have turned it into a mass-market earning app on the App Store. I haven't seen a clear copycat of Reps either. Shift already has comparison pages on its site ("Shift vs Onsite", "Shift vs HighFive") to capture search terms, so the category is getting crowded.

Biggest problem: the supply of tasks isn't in its hands

The complaint in the opening comment is true. In early November, around the 1.1.25 release, Strap tasks disappeared from the app entirely, and quest pay dropped (Facebook group post from Nov 2, 18 comments; several App Store reviews from around Nov 1). The homepage still says "Capture it hands-free with your Reps Strap," and the Nov 6 update, 1.1.26, says only "Bug fixes and performance improvements."

I don't know why they pulled it. Possibly the buyers didn't want enough video, review costs were too high, or they're shifting toward data like receipts and location (inference). Whatever the cause, it shows that how many tasks these apps can offer and how much they pay is set by whoever buys the data, not by the app. Once Strap was pulled, the paid reason to subscribe to Ultra for Strap also fell apart. Reverse-sentiment videos are already on TikTok, e.g. @fineapplewuu: "DO NOT GO ON THIS APP YOUR INFORMATION IS BEING SOLD."

If it were you

C: the funnel is worth studying, but the idea itself isn't worth chasing. Reps took off by using a proven reward-app playbook (small tasks that pay out right away, a wallet balance, withdrawal thresholds, paid UGC, two-tier referrals), on a theme the media had just hyped. Anyone can copy the playbook. What's hard to copy is the buyer behind it: without a robotics company paying by the hour for your videos, there's no money to pay you. Reps itself was cutting tasks in November.

Who it suits: teams that already have data buyers or can land them (people who've done data labeling, crowdsourcing, or research panels), and people who've run reward apps like Freecash or Mistplay and know how to run UGC accounts and referral programs.

Who it's not for: anyone with no B2B channel who expects buyers to show up after they build an app; anyone counting on a "pay first to withdraw" model, which reviewers already call a SCAM and which draws fire from Apple's review team and the press; and anyone targeting markets outside the US. Reps only takes US households, and nobody has checked whether people in other countries will buy household chore videos.

How to do it: find the buyer first, then build the app. Ask robotics and embodied-AI companies that are collecting data whether any specific scenario is short on video (kitchens, repairs, warehouses, households in a particular country), and build supply for that. On the app side, copy Reps: there should be tasks you can do as soon as you install, a visible wallet balance, and gates on high-paying tasks. Don't make users pay before they can get back what they've earned. Setting the gate at "record a set number of qualified video minutes" is enough. Or focus on one type of data nobody has done well, such as receipt-plus-location consumer panel data. The buyers there are brands and research firms, which are more numerous than robotics companies and much steadier.

How to get users: on TikTok, find side-hustle and beer-money style accounts, put a direct link in the bio, and use a shared hashtag template. Get into side-hustle and "AI training" groups on Facebook and Reddit, and post tutorials and earnings screenshots. Launch with two-tier referrals from day one. Use search terms like "Get Paid" in the store title.

Before when: This round of robot-data funding started in 2026, and buyers are paying by the hour now because everyone is short on data. Once a few big collectors (XDOF, MicroAGI) scale up supply, or simulation and synthetic data get good enough, prices will slide (inference). If you're going to build this, lock in buyers before the first half of 2027.

What would show I got it wrong: (1) Reps restores Strap tasks within a month or two and its daily rating growth stays at a hundred or two hundred a day, which would mean buyer demand is steadier than I think, and this should be rated higher; (2) the reverse, if Reps's and Amelia's new daily ratings drop to a few dozen, or Amelia also starts cutting tasks, the buyer side is tightening, and you'd need to change themes, not just copy the playbook; (3) if the three to five robotics companies you approach all say "we already have a supplier," drop it.