Loading page…
Auto-turns blog posts, long videos, and songs into ready-to-publish viral short-form videos
A textbook pivot: users "bending" your product to do something else is the strongest pivot signal there is — so he decisively rebranded and rebuilt.
Fewer bars = easier, cheaper, or faster for an AI-assisted solo builder. Editorial judgments based on the case details.
Tibo acquired Tyframe, a small product that made product promo videos (about $12K/mo at the time). He admits he made the classic post-acquisition mistake: rushing to sell hard instead of validating — "when your main goal becomes pitching, you can no longer hear what users are saying." Then he noticed two things: product videos are a low-frequency need with heavy churn, and users were "bending" the product to do other things. His exact words: "When your product is designed for one specific thing and people keep hacking it into something else, that's usually a very strong signal that you should follow that direction." So he rebranded, bought a new domain, and pivoted into Revid — right at the start of the AI video wave: "every time a new model ships, the product gets more capable, so it's caught fire several times." Today it makes $600K+/mo with a team of 4-5.
Revid did not start from zero in the usual sense — Tibo bought the starting line. He acquired Tyframe, a small product making product promo videos at about $12K a month, and inherited its users. What happened next is the part worth copying: he admits he rushed to sell hard instead of listening, and "when your main goal becomes pitching, you can no longer hear what users are saying." The first genuinely new customers came from watching what the inherited ones actually did with the tool, which was bending a promo-video maker into other jobs — and that is where Revid came from. His standing method for a product with no users is narrow and manual. Find five to ten people who really are the target, reach them through a tweet, a subreddit or an email, and build a relationship with each one instead of collecting feedback. Until a product passes $10K a month the support link points at his own Twitter DMs, which creates a daily flow of users into his inbox and lets him ship a fix five or ten minutes after a complaint arrives.
For Revid the first dollars came with the asset: Tyframe was already at roughly $12K a month when Tibo bought it, so the open question was never whether someone would pay but whether they would keep paying — promo videos proved to be a low-frequency need with heavy churn. His pricing method runs backwards from the number. He decides first that he wants a use case where he can reliably charge $50-100 a month, then shapes the product to justify it: at that price nobody haggles before subscribing, customer quality is high and churn is low, whereas $9 a month "makes you look like a cheap product" and buys a pile of DMs from people paying ten dollars. The comparison does the selling for him, since human editors charge several hundred dollars to cut one interview into shorts. Across the whole portfolio he reports about 50,000 paying customers against roughly $700K a month, near $14 each, so the $50-100 band describes the tier he designs toward rather than the blended average a mass-market video tool lands on.
Tibo's complete playbook (validation / pricing / churn red lines / SEO matrix) is the most systematic in this entire dataset — use it as a checklist.
Paste into Claude Code / Codex and get a working version of this product end-to-end.
Two production-grade prompts per idea — a full build spec and an SEO growth plan — plus 10 tailored build specs a week. Members only.
Join the Unicorn Club — $5/moCancel anytime, in one clickEvery claim traces back to a listed source — open any of them and check the number yourself.
Data credibility: First-person founder interview. All revenue figures are founder-reported or estimated by the video creator; none are audited.
Two things made Revid's growth self-feeding. The first is the SEO matrix: over a hundred mini tools on the site, one page per use case — audio to video, blog to video, long video to shorts — because "every tool is a chance to rank," and a page you have to actually use resists being replaced by an AI summary in a way an article does not. His keyword research is crude on purpose: google the head term, then treat every result on that page as a page worth building. When he could not outrank competitors with "video" in their domain names, he dropped the head terms entirely and went all-in on niche queries. The second is auto mode — connect a channel or a blog and videos ship automatically whenever new content appears, which turns a tool people open when they remember it into a subscription that runs without them. That is the churn fix, and he insists on the order: get monthly churn under 20% before spending anything on acquisition, because at 40% users renew twice and no amount of bought traffic breaks the ceiling.
The long quiet stretch came before any of this — "I failed for 5 years because I didn't know this" — and the way out was changing the unit of time. For four months he shipped one new product a week: landing page first, promote it on X, watch for a response, and with no response move to the next one on Monday. Nine died and the tenth took off. His arithmetic is that a 90% failure rate is survivable only when an attempt costs a week, because at a year per attempt the same odds cost nine years. The judgement rule is deliberately harsh: revenue is the only signal that counts, and warm feedback with no payment and no retention is how founders talk themselves into another year. The same test governs a product already live, through churn. Under 20% a month is worth continuing, while 40% means the product cannot hold people and the answer is to pivot rather than to market harder. Applied to Revid, that reading is what turned a promo-video tool with a low-frequency need into a different product entirely.
In 2024-25, every AI video model could only generate little 5-second clips — nobody could stitch them together into anything coherent. Revid's core logic is automating that stitching: cut long content into scenes, use its in-house engine to assemble multiple 5-second clips, and keep characters, scenes, and environments consistent throughout. Usage is dead simple: drop in a Spotify link and a music video comes out; upload a photo and "you" sing in the video; the strongest feature is auto mode — connect your channel or blog, and whenever new content ships, Revid automatically turns it into a video. "Talk to video editors — they charge several hundred to cut one interview short. If AI handles the whole thing automatically, that's beyond their wildest dreams."
The site has 100+ mini tools, each use case with its own page: audio-to-video, blog-to-video, long-video-to-shorts... "Every tool is a chance to rank in search engines," because for a concrete need like "audio to video," Google can hardly serve an AI summary that replaces the tool — "this isn't just content, it's a tool you have to actually use." His scrappy keyword research method: search Google for "AI video tools" — "every result that comes up is a keyword worth fighting to rank for." Domain relevance genuinely matters: unable to beat rivals with "video" in their domains, he abandoned the big "AI video" head terms entirely and went after niche terms — "I've done pretty well on that front." He also points out that Pieter Levels' Photo AI runs the same play: masses of keyword-dedicated landing pages like "AI Easter photos" — "the guy has a massive following, and he still spends serious time building search-traffic pages."
Pricing: he works in reverse — first deciding "I want a use case where I can reliably charge $50-100 a month," then shaping the product around that price. His reasoning: "At that price it's not a lot — nobody asks me for discounts or haggles before subscribing; but it's not cheap either, so customer quality is high and churn is low. $9/month works too, but it makes you look like a cheap product, and you face a pile of DMs complaining — they pay you ten dollars a month; is replying to every single one really worth it?" Retention: monthly churn under 20% is decent and worth continuing; 40% means users only renew for two months — "your product simply can't hold onto users." He explains the Max MRR concept: churn and acquisition together make revenue grow along an S-curve toward some ceiling; when churn is high, no amount of bought traffic breaks through — "which is why pivoting is far better: get churn below 20% first, then think about acquisition." One addendum: five years ago people only respected products with churn under 5%; now that the AI crowd hops between products so fast, 20% is the new baseline.
Before his first successful product (the one that got acquired), he went through a crazy four months: shipping one new product every week — build the landing page first, promote it on Twitter, watch the market's response; no response, move straight on to the next; only with a response would he seriously build it into a product. Nine flopped; the tenth took off. His reflection: "I always thought quality mattered most. Later I realized that building with an emphasis on quantity ends up producing better quality — picking the best out of a pile of products is far more reliable than grinding away on a single one." Validation counts revenue only: "The people on Twitter bragging about raising X million or hitting some revenue number are talking nonsense. It's too easy to fool yourself into believing the product has made it because you got some validation-flavored feedback — but with no revenue and no retention, it's very hard to succeed."
Tools: Cursor for writing code; GPT-5.4 for complex logic, Gemini 3.1 Pro for UI ("two weeks ago my favorite was still a different model — it changes that fast"). Process: have the AI produce a spec first; he scraps and regenerates about 50% of the specs — "steering the AI mid-course is too complicated. I just go back to the setup, tweak it, and run it again until the spec it gives me feels solid." The stroke of genius: before implementing, have the AI write a prompt for Nano Banana to generate a "preview image of the UI once the feature is done" — "one glance at the image and I know which parts I like and which I don't — far better than staring at dry text." He also recommends Pencil (a tool that produces design mockups right inside Cursor). On new ideas: "I used to have a head full of ideas I'd never finish building. Now my idea list is empty — the moment I have one, it's done in 10 minutes. The hardest parts have become coming up with new ideas and distribution."
Support: keyboard shortcuts — typing "refund" auto-generates a complete refund reply and routes the request to the person responsible; "the moment I start repeating something, I always find a way to automate it." Once the product grew he hired someone dedicated to support, but he keeps his DM channel open — "so I always have a read on what users are up to." Team: deliberately kept ultra-lean — "one person today can do the work 20 people used to do; with few people, communication is simple, adopting new tools is fast, and pivoting is easy. Back when a 20-person team wanted to turn the ship, it took enormous effort convincing them one by one — overnight you'd go from great boss to shortsighted founder. I never want to suffer that again."
The AI short-video space itself is already crowded — copying the product outright comes at a discount. But Tibo's methodology (pivot signals, validation cadence, pricing, churn red lines, the SEO matrix, the dev workflow) is the most systematic and actionable in this entire dataset — treat it as a checklist you reread before starting any project. His goal-setting line is a gift too: "Don't force yourself toward $1M months. A steady $10K a month is already a very comfortable life — that genuinely counts as an enormous success."
That was one of the 50 breakdowns open to everyone. The Unicorn Club opens the other 1,285 at the same depth — plus pain-point filters and 10 build specs a week.
Join the Unicorn Club — $5/moCancel anytime, in one click