The one-glance read on who they are and how they grow. Each point is verifiable from the receipts above.
An AI stem-separation and practice app for musicians (vocal/instrument isolation, pitch, tempo, chord detection) with 70M registered users reported as of December 2025, plus a B2B audio-AI platform, Music.AI, that opens the same underlying tech to enterprise developers.
Organic-led, with a heavy, sustained paid-acquisition layer running underneath rather than driving it.
≈3.95M monthly visits (+23.4% over the last 3 months) as of 2026-07-11, a 236,547-follower owned Instagram, and a Google ad library of roughly 2,000 units with one active ad running 1,088 days as of 2026-07-11.
As of 2026-07-11, the newest Meta creative routes clicks straight to moises.ai instead of an app-store listing, while Google is testing fresh image and video formats first shown between July 1 and July 7, 2026.
Most of Moises's organic pull runs through a Portuguese-language homepage capturing 71.8% of its organic-content traffic, with a multi-language (Portuguese, Spanish, Japanese, English) paid layer stacked on top rather than a single global campaign.
The order the channels came online. Sequence is strategy: what they did first, and what they layered on once demand existed.
Estimated demand, the channel split behind it, and the keywords and referrers doing the work. Directional modeling, not audited analytics.
| Keyword | Volume | Weight | CPC |
|---|---|---|---|
| moises | 330k | $0.22 | |
| moises ai | 95k | $0.29 | |
| moises app | 21k | $0.27 | |
| moise | 15k | $0.33 | |
| moises studio | 4.8k | $1.13 |
Moises pulls an estimated ≈3.95M monthly visits, up 23.4% over the last 3 months as of 2026-07-11. Direct traffic leads at 44.9%, a level consistent with a habitual-use utility app that gets bookmarked or reopened rather than re-discovered, followed by organic search at 22.3% and paid search at 12.1%. Social organic (9.1%) and display ads (4%) round out the measurable mix; the rest, paid social, generative-AI referral, referrals, and email, together totals about 7.5%.
The keywords they rank and bid for skew almost entirely branded: "moises" carries ≈330,000 monthly searches, "moises ai" ≈95,000, and "moises app" ≈20,900, meaning most of that organic-search share is people searching for the product by name rather than category terms like "vocal remover." Referring sites include similarsites.com, a tool-comparison directory that suggests aggregator-driven discovery, vocalremover.org, which also shows up as a direct competitor, implying some users bounce between the two free tools, and stripe.com, most likely a checkout-flow artifact rather than genuine external content. Named competitors span direct stem-separation rivals like lalal.ai and vocalremover.org, and suno.com, an AI music-generation tool in an adjacent category, which lines up with the Reddit thread comparing Moises's newer AI Studio, which generates instrumentals that adapt to a user's own track, to Suno.
The specific pages earning their organic search traffic, and the pattern behind why they rank. Adapt the format, not the topic.
The single biggest organic asset is the Portuguese-language homepage (moises.ai/pt/), which alone earns 71.8% of the pages tracked in this data, followed by a dedicated vocal-remover feature page capturing tool-intent search at 7.5%. Behind those sit Portuguese-language blog posts on "how to sing better" and "how to write a song," plus a platform-specific "Moises for PC" page. The pattern is a Brazil-first content strategy: one dominant branded/navigational hub, one single-feature landing page built to catch a specific search intent, and how-to blog content that captures top-of-funnel musician queries rather than competing on broad "AI music" terms. It likely ranks well because the brand already owns the top position for its own name and app-name variants, and the feature and blog pages extend that authority into adjacent, lower-competition queries.
Not traffic share. How much weight the growth system actually puts on each channel, with a one-line read on the role it plays.
Open roles read like a roadmap: the functions they’re staffing show where the company is investing next and what stage it’s at.
For founder-led SaaS the breakdown shifts from ads to traction: where the first users came from, how the founder grows it in the open, and the compounding organic surface.
Public traction data for Moises's launch is thin: what survives is the founder's own account of two early product milestones and a scraped description of the product's earliest prototype, not a documented traction spike.
The earliest public trace of Moises describes a browser-based interface built on top of a newly published third-party audio-separation engine, the same technical seed later folded into Moises's own stem-separation product.
Geraldo Ramos, one of Moises's co-founders and its CEO (the rest of the founding team isn't independently verified in this evidence), tweeted "Just launched on @ProductHunt" on 2019-06-07 and "We just launched our native apps!" on 2020-12-01, each from the personal account covered above, though no upvote, comment, or install counts survive alongside either post.
There is no record here of a single traction-inflecting moment, no upvote count, no press pickup, no viral thread, so the growth that followed most likely accumulated through app-store search and word of mouth rather than one documented launch spike.
The founder's own account carries far less GTM weight than the brand's paid and organic machine, functioning as an occasional milestone-announcement channel rather than a build-in-public engine.
Geraldo Ramos posts from a personal X account with 1,147 followers and 5,785 total posts covered above, but only a handful surface as product-relevant, spaced years apart, including the 2019 Product Hunt note and the 2020 native-app announcement.
His highest-engagement tweet, "We just released a new instrument-based AI music model at @moises_ai," from 2025-08-20, drew 38 likes and 7,313 views, modest next to the app's registered-user base, showing the account amplifies shipped features rather than cultivating an audience.
The evidence shows no founder posts on LinkedIn or Reddit, so Moises's open-in-public motion doesn't extend past X, meaning a reader hoping to copy a "founder as growth channel" playbook won't find a worked example here.
With the founder's account this quiet, the owned Instagram account and the Charlie Puth campaign covered above do the public storytelling instead, alongside the creator UGC covered under TikTok.
Moises's organic-content footprint is narrow but concentrated, and the real strategic layer sits in language targeting and brand-term defense rather than a sprawling content library.
The PT homepage covered above sits alongside an open Creator Content Lead (Brazil) role, a deliberate bet on producing more creator content specifically for the market that already anchors their organic traffic.
Only five pages carry meaningful organic weight per the data covered above, a concentrated bet on one homepage, one feature page, and two how-to posts rather than a sprawling blog archive, which implies the flywheel here rewards owning a handful of high-intent pages well over publishing broadly.
With the keywords they bid on skewing almost entirely branded, per the mix covered above, the multi-year-running Google text ad "AI Audio Separation - Separate Instruments w/ Moises," a problem-agitation hook, functions as defensive brand-term capture, making sure a searcher typing the product's own name lands on a Moises-controlled result rather than a category-demand play.
The Meta creative covered above includes curiosity-hook demos like "¿Te puedo contar un secreto?" ("Can I tell you a secret?"), which delays the product reveal before showing the app, doing the top-of-funnel discovery work that branded search ads structurally can't.
This founder-led SaaS also runs paid acquisition. Here are the live ads doing the work, each with the X-ray and a play you can adapt.
Why it works. Withholding what's inside an unmarked case in a moody, disco-lit studio makes viewers wait through the setup to see the product revealed.
Sensory Synchronization (0:03-0:18) Delay the product reveal behind an unexplained prop in a setting your audience already recognizes, and let the reveal happen mid-video.
Why it works. Opens with a direct personal question in the viewer's own language before revealing the product, using a bookshelf and amplifier to set a musician's room scene.
Real-time sensory demonstration (0:03-0:25). Open with a direct, personal question to camera in your audience's own language before you show the product.
Why it works. Leads with an unqualified best-in-category claim in Spanish, then earns it by pulling the phone out of a container of drumsticks, a prop from a musician's world.
Real-time audio isolation (0:03-0:07) Pair a superlative claim in your headline with a physical prop from your audience's world so the visual backs up the words.
Why it works. Runs the same best-in-category claim in Portuguese for the Brazilian market, using the drumsticks-to-phone reveal to root the app in a musician's world.
Real-time audio-visual synchronization (0:05-0:14) Reuse a proven claim-plus-reveal structure across markets, translating the claim natively instead of running one language everywhere.
Why it works. Placing the phone directly on a drum cymbal puts the app in the physical spot a drummer would actually use it, rather than an empty background.
POV framing on a drum kit (0:00) Shoot your product resting on the real object or surface your customer uses it next to, not on a plain table or hand.
Why it works. Declaring the search already resolved speaks directly to someone actively searching, paired with a tactile drumstick visual for niche relevance.
Real-time audio manipulation (0:03-0:25) For a searched-for category, open with 'you just found...' language rather than introducing the category from scratch.
The channels are not separate. They are one system where each stage feeds the next. Here is the read, then the plays to run tomorrow.
The loop runs on brand-search defense and paid-social installs feeding a Brazil-anchored content base, with no visible founder-audience or broad-content layer to compound discovery beyond the brand's own name.
Multi-language paid social and video ads covered above route users straight to app-store install pages (play.google.com, itunes.apple.com dominate the destinations), functioning as direct-to-install demand capture rather than top-of-funnel education, though the newest Meta creative is now testing a direct-to-site route to moises.ai instead.
Reddit users describe a Premium subscription and a newer Pro tier that isolates individual drum components (kick, snare, hihat, toms), the tier split visible only through user testimony rather than Moises's login-gated pricing page.
The PT homepage dominance covered above, the Portuguese/Spanish/Japanese paid layer, and the new Brazil-focused creator hire stack into one reinforcing regional advantage rather than a global content machine.
With founder posting this sparse and organic keywords running almost entirely branded, there's no visible discovery engine for people who aren't already searching "moises," leaving paid installs and brand-search defense to carry weight that organic content and a founder audience would otherwise share.
free-to-paid conversion rate, churn, exact subscription pricing, CAC.
The proofMoises's earliest public trace, covered above, describes a browser-based interface built on someone else's published audio-separation research, not a custom model trained from scratch.
The adaptationFind a published open research model or paper in your category, ship a simple interface around it within days instead of building custom infrastructure, and put it in front of people for free. The mechanism: wrapping proven research in a usable interface tests real demand at near-zero technical cost, so you defer the "build our own tech" decision until you know anyone wants it.
Cost: $0 · Time to signal: days · Works pre-PMF: yes.
The proofGeraldo Ramos announced the Product Hunt launch and the native-app release from his personal account, covered above, when it had a small following, tying each post to a concrete shipped feature rather than a vague update.
The adaptationStart posting every real shipped milestone, not progress updates, from your own account before you have an audience worth mentioning. State what shipped, why it matters, and a link. First step: write down the next feature you ship and post it the day it goes live, nothing more polished than that.
Cost: $0 · Time to signal: days per post, months to build credibility · Works pre-PMF: yes.
The proofThe creator UGC roster covered above includes a bassist (kingaglyk) whose organic, unpaid demo of the app drew 3,657 engagements, alongside other musicians posting unscripted use cases.
The adaptationIdentify five to ten real customers active on social media in your niche, give them free access or a small gift, and ask for an unscripted "how I use this" post rather than a scripted ad. First step: DM the most engaged users already talking about your product and offer them early access to a new feature in exchange for an honest post.
Cost: $0-under $500 · Time to signal: weeks · Works pre-PMF: yes, provided you already have a handful of genuine early users willing to post.
The proofThe vocal-remover feature page earns 7.5% of the tracked organic pages on its own, separate from the branded homepage.
The adaptationPick one specific "how do I do X with your product" query your users actually search, tied to a single feature rather than your brand name, and build one clean page whose title, headline, and copy match that exact phrase. First step: pull your top feature-related support question and turn it into a dedicated landing page this month.
Cost: $0 · Time to signal: weeks · Works pre-PMF: yes.
The proofThe Charlie Puth Chief Music Officer campaign covered above turned one collaborator into a recurring content source, including a remix contest and reels reaching up to 40,890 engagements.
The adaptationIdentify a respected figure in your niche, not necessarily a celebrity, and offer them a real ongoing title or advisory credit with creative input instead of a flat sponsorship fee, then build a recurring mechanic (a contest, a monthly feature) around them so the relationship produces content on a cadence. The mechanism: an ongoing named role gives the collaborator a reason to keep creating, since their name is now attached to the product's roadmap, unlike a single paid post.
Cost: dedicated budget or hire · Time to signal: months · Works pre-PMF: no, this requires an existing product and audience substantial enough to make the collaboration credible. Not transferable at an earlier stage: the multi-language paid ad library, the four-language creative localization, and the celebrity Chief Music Officer role all assume a funded team and existing scale; a pre-PMF founder should adapt the underlying mechanics in Plays 1-4, not attempt the paid or celebrity motion directly.
Every morning we take one company that is actually growing and break down where its customers come from: the ads still running after a year, the channel doing the real work, and the play you can run this week.
Systemaic · directional intelligence. Traffic, spend, and reach figures are SimilarWeb-style estimates and qualitative reads of public data, not audited numbers. Built on real public receipts.
Live ad libraries: meta ad library · linkedin ad library · google ad library
Launch archives: Hacker News: Browser-based interface for the new Deezer’s audio separatio
Traffic, spend, and revenue figures are estimates as noted in the report; the links above are the primary public artifacts.