Cedric Roberge
Founder of Pep AI, a peptide and GLP-1 tracking app, college student
Where to hear and read about Cedric Roberge1 episode or test
Cedric Roberge is a college student who founded Pep AI, a peptide and GLP-1 tracking app that reached $100,000 in revenue in 70 days. The most useful lesson is to sign nearly every creator in a trending niche on small performance-based deals, then turn their winning posts into cheap paid ads.
None of them are doing influencer marketing. They're only running ads.
I speedran an ai app to $100k in 70 days.
Cedric Roberge: the story and the lessons
1How Cedric Roberge started Pep AI
Cedric built and launched Pep AI as a college student in about a month. He began building in late January and launched at the end of February. The app reached $100,000 in revenue in 70 days. He built the first version himself.
He says he built the app in under three months.
He saw peptide trackers as an empty, neglected market. Peptide calculator apps had existed for years with only a couple of reviews and no marketing. GLP-1 trackers like Ozempic apps were already doing well, but peptides had nothing.
Peptides are a wider market than GLPs, which gives creators more content angles. Cedric says his app helped push the peptide trend itself.
Running a Fortnite esports team taught him the creator system he reused. In high school his team had hundreds of thousands of followers across platforms. He ran 30 to 50 players on Discord, collected their best clips and sent them to an editor.
He studied Zach King, Lele Pons and King Bach for influencer marketing. The affiliate deal was not planned; it came to him later, and he tested many things early.
One investor paid for marketing and equity brought in the developers. A fraternity neighbor who had run dropshipping for years invested after one pitch call, though Cedric barely knew him. All of the money went to marketing.
Two developers from a university founders club joined about a month in for a share of the company. They now run all development, which frees Cedric for marketing, product, pricing and ads.
2How Cedric signed almost every peptide creator
Cedric copied the channel rival GLP apps ignored. Rival GLP apps ran only ads, so he studied apps that won on creators, such as CaliAI and Quitter, and signed nearly every consistent peptide creator. Ads only started in the third month.
Revenue went from $30,000 to $50,000 to about $80,000 a month on roughly $10,000 to $25,000 of monthly marketing spend. About 100 creators post.
He starts every creator on a $200 trial month. The trial is four videos and four stories plus a bio link. Pay follows views and engagement, not follower count, and about 5,000 views per video earns roughly $200.
A strong month raises the fee and can lead to a 3-month deal. An okay month keeps it the same and a weak one lowers it. Bigger creators earn up to $4,000 a month.
He bet over $20,000 on one large fitness creator. The creator had worked with only one other app, where he was the number one marketing channel. He starts posting next month, and first-month spend was $10,000 to $15,000.
Cedric calls the deal a risk. It came only after the smaller retainer system had proven itself.
A 20% one-time affiliate cut makes creators work harder. Almost every creator has an affiliate deal, usually 20% of subscription sales, paid once. Many post more than asked and put more effort into each post.
New features also become a pitch to creators, since a feature people like means more sales on their code. Cedric can see all subscription sales, which makes paying the split workable.
Tracked links and creator codes let him price each creator. AppsFlyer OneLinks log click, download and purchase per creator, and codes at signup show who referred each user. Profit is checked per creator, and weak ones get lower prices.
The link opens the App Store or Google Play depending on the phone.
Bio links bring downloads even when a creator posts nothing. One creator who stopped posting for 2 weeks still drove a ton of downloads. Followers of informational peptide creators click bio links to learn more and find the app there.
As Cedric puts it, "he didn't even make a single post."
He gives creators examples, not scripts, and reviews the first video. A Google Doc of best-performing posts says to recreate them or do your own thing. Ethan's slideshows are the top performer, and Moose's 19-second joke video got 400,000 views.
New creators send their first video to a group chat before posting. A few were only about the creator's transformation and were redone, and the redos were great.
He vets creators on comments after paying for bought views. One signed creator had 120,000 views in an hour but no comments and no sign-ups. He now checks comments and real engagement before signing anyone.
"The biggest thing is looking for real engagement, um especially comments."
Daily contact keeps creators invested without formal view bonuses. The team texts nearly every creator daily and calls them. New creators get lifetime app access before any deal.
Some answer follower questions and go on TikTok Live for the app unprompted. Cedric calls them "the backbone to our success."
3How Cedric runs paid ads and delegates the machine
He runs only proven organic winners as Meta ads. Scripts copied from other GLP apps failed and cost money. Winning TikTok and Instagram posts performed well on Meta, with budget raised weekly on the winners.
Cost per download is about $2.50 against a $5 rule of thumb. Spend is about $400 a day on Meta plus $150 on Apple Search Ads.
A cluttered static ad he dislikes became his best performer. The ads lead proposed a square graphic with about eight pieces of information. It gets under $2 per download and was made in about 2 minutes with ChatGPT.
Cedric says it is "way too cluttered" but it performs, so he now orders more variants, including cleaner ones for men and women. He cannot explain why it wins.
He buys paid ads on Meta because TikTok rejects peptide content. TikTok keeps rejecting Spark ads and removing videos because peptide content reads as medical advice. Rival GLP apps in the App Store top 100 for Health & Fitness use only Meta ads.
Pep AI ranks about 250 and aims for the top 100.
Outreach runs at volume from one fixed opening line. The team sends over 100 DMs a day and over 1,000 a week. The first line is "paid promo / affiliate deal?", and replies move to WhatsApp or iMessage and then a call.
Cedric skips accounts with about 200 followers and no views. He takes calls with bigger creators and a marketing manager takes the rest.
He built the process first, then taught friends to run it. Marketing managers handle replies, numbers, the call, the deal, the contract, payment and code setup. The team is four outreach hires, two marketing managers, two developers, a CFO and a product person.
Hires follow a standard procedure instead of inventing one. As Cedric says, "I would just teach them."
4How Cedric shapes the Pep AI product
He shortens onboarding for under-25s and lengthens it for everyone else. About 60% of users are over 25. Under-25s buy less and drop out of long flows, so they get key points and disclaimers. Older users get a long flow covering nutrition, water and sleep, plus a personalized nutrition plan.
An onboarding mistake had earlier lost a meaningful share of sign-ups before the paywall.
He keeps the paywall low in commitment and tests cheaper prices. The yearly plan has a 3-day free trial, at $40 a year or $10 a month. He is working on going cheaper, and prices are always being tested.
A flash sale borrowed from e-commerce gave his biggest revenue day.
He frames the app as user-input only to avoid medical advice. All data is entered by users and the app gives no recommendations. The AI chat answers only simple questions such as travelling with a GLP.
He avoided the syringe calculator for a long time because it sounded like medical advice. After talking to attorneys he shipped it, and it drove a ton of traffic.
He builds what users ask for and cuts what they ignore. Cycle scheduling and the syringe calculator came from feedback. The AI body scan and bloodwork features were removed, and achievements were cut to simple streaks.
Feedback comes from in-app forms, email, reviews, DMs and calls with creators. The team ships two to three updates a week.
I speedran an ai app to $100k in 70 days.
Cedric built a peptide and GLP-1 tracking app and used a creator-first retainer plus affiliate system to reach $100,000 in revenue in 70 days. He explains how he sources and pays creators, turns organic winners into Meta ads, and tailors onboarding by age.
Cedric built and launched Pep AI as a college student in about a month. He began building in late January and launched at the end of February. The app reached $100,000 in revenue in 70 days. He built the first version himself.
He says he built the app in under three months.
He saw peptide trackers as an empty, neglected market. Peptide calculator apps had existed for years with only a couple of reviews and no marketing. GLP-1 trackers like Ozempic apps were already doing well, but peptides had nothing.
Peptides are a wider market than GLPs, which gives creators more content angles. Cedric says his app helped push the peptide trend itself.
Running a Fortnite esports team taught him the creator system he reused. In high school his team had hundreds of thousands of followers across platforms. He ran 30 to 50 players on Discord, collected their best clips and sent them to an editor.
He studied Zach King, Lele Pons and King Bach for influencer marketing. The affiliate deal was not planned; it came to him later, and he tested many things early.
One investor paid for marketing and equity brought in the developers. A fraternity neighbor who had run dropshipping for years invested after one pitch call, though Cedric barely knew him. All of the money went to marketing.
Two developers from a university founders club joined about a month in for a share of the company. They now run all development, which frees Cedric for marketing, product, pricing and ads.
Cedric copied the channel rival GLP apps ignored. Rival GLP apps ran only ads, so he studied apps that won on creators, such as CaliAI and Quitter, and signed nearly every consistent peptide creator. Ads only started in the third month.
Revenue went from $30,000 to $50,000 to about $80,000 a month on roughly $10,000 to $25,000 of monthly marketing spend. About 100 creators post.
He starts every creator on a $200 trial month. The trial is four videos and four stories plus a bio link. Pay follows views and engagement, not follower count, and about 5,000 views per video earns roughly $200.
A strong month raises the fee and can lead to a 3-month deal. An okay month keeps it the same and a weak one lowers it. Bigger creators earn up to $4,000 a month.
He bet over $20,000 on one large fitness creator. The creator had worked with only one other app, where he was the number one marketing channel. He starts posting next month, and first-month spend was $10,000 to $15,000.
Cedric calls the deal a risk. It came only after the smaller retainer system had proven itself.
A 20% one-time affiliate cut makes creators work harder. Almost every creator has an affiliate deal, usually 20% of subscription sales, paid once. Many post more than asked and put more effort into each post.
New features also become a pitch to creators, since a feature people like means more sales on their code. Cedric can see all subscription sales, which makes paying the split workable.
Tracked links and creator codes let him price each creator. AppsFlyer OneLinks log click, download and purchase per creator, and codes at signup show who referred each user. Profit is checked per creator, and weak ones get lower prices.
The link opens the App Store or Google Play depending on the phone.
Bio links bring downloads even when a creator posts nothing. One creator who stopped posting for 2 weeks still drove a ton of downloads. Followers of informational peptide creators click bio links to learn more and find the app there.
As Cedric puts it, "he didn't even make a single post."
He gives creators examples, not scripts, and reviews the first video. A Google Doc of best-performing posts says to recreate them or do your own thing. Ethan's slideshows are the top performer, and Moose's 19-second joke video got 400,000 views.
New creators send their first video to a group chat before posting. A few were only about the creator's transformation and were redone, and the redos were great.
He vets creators on comments after paying for bought views. One signed creator had 120,000 views in an hour but no comments and no sign-ups. He now checks comments and real engagement before signing anyone.
"The biggest thing is looking for real engagement, um especially comments."
Daily contact keeps creators invested without formal view bonuses. The team texts nearly every creator daily and calls them. New creators get lifetime app access before any deal.
Some answer follower questions and go on TikTok Live for the app unprompted. Cedric calls them "the backbone to our success."
He runs only proven organic winners as Meta ads. Scripts copied from other GLP apps failed and cost money. Winning TikTok and Instagram posts performed well on Meta, with budget raised weekly on the winners.
Cost per download is about $2.50 against a $5 rule of thumb. Spend is about $400 a day on Meta plus $150 on Apple Search Ads.
A cluttered static ad he dislikes became his best performer. The ads lead proposed a square graphic with about eight pieces of information. It gets under $2 per download and was made in about 2 minutes with ChatGPT.
Cedric says it is "way too cluttered" but it performs, so he now orders more variants, including cleaner ones for men and women. He cannot explain why it wins.
He buys paid ads on Meta because TikTok rejects peptide content. TikTok keeps rejecting Spark ads and removing videos because peptide content reads as medical advice. Rival GLP apps in the App Store top 100 for Health & Fitness use only Meta ads.
Pep AI ranks about 250 and aims for the top 100.
Outreach runs at volume from one fixed opening line. The team sends over 100 DMs a day and over 1,000 a week. The first line is "paid promo / affiliate deal?", and replies move to WhatsApp or iMessage and then a call.
Cedric skips accounts with about 200 followers and no views. He takes calls with bigger creators and a marketing manager takes the rest.
He built the process first, then taught friends to run it. Marketing managers handle replies, numbers, the call, the deal, the contract, payment and code setup. The team is four outreach hires, two marketing managers, two developers, a CFO and a product person.
Hires follow a standard procedure instead of inventing one. As Cedric says, "I would just teach them."
He shortens onboarding for under-25s and lengthens it for everyone else. About 60% of users are over 25. Under-25s buy less and drop out of long flows, so they get key points and disclaimers. Older users get a long flow covering nutrition, water and sleep, plus a personalized nutrition plan.
An onboarding mistake had earlier lost a meaningful share of sign-ups before the paywall.
He keeps the paywall low in commitment and tests cheaper prices. The yearly plan has a 3-day free trial, at $40 a year or $10 a month. He is working on going cheaper, and prices are always being tested.
A flash sale borrowed from e-commerce gave his biggest revenue day.
He frames the app as user-input only to avoid medical advice. All data is entered by users and the app gives no recommendations. The AI chat answers only simple questions such as travelling with a GLP.
He avoided the syringe calculator for a long time because it sounded like medical advice. After talking to attorneys he shipped it, and it drove a ton of traffic.
He builds what users ask for and cuts what they ignore. Cycle scheduling and the syringe calculator came from feedback. The AI body scan and bloodwork features were removed, and achievements were cut to simple streaks.
Feedback comes from in-app forms, email, reviews, DMs and calls with creators. The team ships two to three updates a week.