I'm Fuljens Henry. I produce original Haitian kompa and gouyad as LAN8CHO — seven records, four voices, all of it written, arranged and directed by me. Now I build custom music for brands, artists and creators.
Vocals are AI-generated. Everything around them is production.
Seven records, streaming everywhere. Pick a lane — Kompa Lanmou for the modern, late-night side. Konpa Direk for 1970s Haiti with a full band in one room.
Twa Jou, Hein? · Mwen Te Di W? · Prete M Chemiz La · Dousman, Chérie · Chèz Vid · No Me Des Tensión · Yon Ti Wonm · Lòt Bò Dlo · Tann Jou Pa W
Lè Chat Pa La · Dèyè Mòn · Marie · Bouch Pa Gen Baryè · Ay, Mamita · Wòch Nan Solèy · Yo Di...
Streaming now. The full-album video lands on the channel shortly.
Some Kalix, some tracks with no named artist, and Carlene — the first female voice built for the project. Everything after this is a designed album under a named voice.
Also released: Pran Feeling (Live) and the single Poko Efase'l.
Not reach for its own sake — a specific audience, in specific places, playing the catalog on repeat.
The United States, France, Canada and the Dominican Republic — concentrated in Chicago, Brooklyn, Miami, Fort Lauderdale, Orlando, Montreal, Paris and Port-au-Prince.
Four platforms, measured separately, return the same map: the Haitian diaspora and the francophone markets around it. The TikTok audience for Stay Home was 90% United States, and its top five cities are that same diaspora corridor.
On the streaming platforms, roughly 80% aged 25–54, with almost no audience under 25, and men and women split close to evenly.
TikTok is a different room: 58% under 35 and 62% women. Same songs, younger crowd — worth knowing before deciding where a release goes first.
3,868 saves and 2,883 playlist adds — this is a catalog people file and come back to, not one they sample.
All of it earned from the artist page and listeners’ own libraries. Zero editorial playlist placements. Nobody put this in front of anyone — they went looking for it.
Apple Music for Artists, Spotify for Artists, YouTube Studio and TikTok Analytics · September 2026
You bring a brief, an occasion, a brand, a feeling. I bring the concept, the lyrics, the arrangement, the vocal direction and the finished master. In your genre. In your language.
Start a projectOriginal music built to a spec — a launch anthem, a campaign bed, a theme, a jingle that doesn't sound like a jingle. Delivered in every length you need, plus the instrumental.
Discuss a briefThe full LAN8CHO method applied to your project: a written manifest, a vocal persona that holds across an entire catalog, and a production standard that makes twelve tracks sound like one artist instead of twelve sessions.
Talk it throughAnyone can type a prompt and get a song. That's not what this is. The tool is Suno. The value is direction — knowing what to ask for, what to reject, and how to make eighteen tracks sound like one artist instead of eighteen prompts.
Every album begins as a written document — track list, language, instrumentation, era, influence, and the emotional arc from track one to the close. Nothing is generated until the record is designed.
Recurring characters, established facts, callbacks between albums. Tracked in writing, so album three can collect a debt planted on album one.
A per-album table controlling variety — which instrument leads which track, which cast appears, which language. No instrument leads twice. This is what stops a record being one backing track twelve times.
Instrumentals first, judged and rejected until one is worth keeping. Vocals second, over a locked arrangement, with tested settings rather than guesswork.
The breakthrough that made the voices work: describe an emotional performance, not a vocal texture. Texture plus negatives gives you rasp with no identity.
Once a take is right it's saved as a persona. Text descriptions drift; personas hold — which is how a voice survives a whole catalog.
A running list of what breaks and the fix — pronunciation control in mixed-language lyrics, crowd bleed, genre drift mid-track, lyrics entering early. Solved once, never again.
Too clean, correct but lifeless, drifted on the cover — each has a known adjustment. Guessing is what makes AI music sound like AI music.
Finished audio, artwork, full-album visualizer, vertical cuts with burned-in captions, thumbnails and complete platform metadata.
Eight of these steps happen before anything gets generated — and the ninth is throwing most of it away. What reaches you is what survived.
LAN8CHO is one artist name containing several fully specified vocal personas — each with its own era, band, instrumentation, vocal technique and recurring cast.
Modern kompa and gouyad with R&B, trap-soul and electronic production absorbed underneath. Longing, obsession, jealousy, late-night vulnerability. Danceable while it hurts. Haitian first — everything else is seasoning.
Twoubadou and compas direct. Small dry band, one room, behind the beat. An unreliable narrator explaining himself to a neighbourhood that already knows — with a recurring cast that carries across both records, and a dead bandmate whose tenor sax now plays his ghost.
Full horn section, frontman-driven. Sings ahead of the beat at the top of his range and stays there. Every song happens at a gig. He addresses the room; he never confesses to a listener.
A female voice built on Cho Sessions Vol. 1, before the single-persona album structure existed. Proof the method builds a voice to spec rather than one man in three moods.
| Ti Sonson | Zoklo | |
|---|---|---|
| Era | 1969–1972 | 1979–1983 |
| Band | Small, dry, one room | Full horn section, loud |
| Placement | Behind the beat | Ahead of the beat |
| Space | Leaves it | Fills it; horns cut him off |
| Stance | Explains himself | Explains nothing |
| Person | First — narrates to a listener | Second — addresses the room |
| Territory | The yard, the porch, the block | The gig, the venue, the night |
They come from the same tool, and they don't sound like the same person — that gap is the entire job. Timbre isn't what tells singers apart. Rhythmic placement, phrase length, effort, breathing and the band around the voice are. Those are production decisions, not prompts.
Every LAN8CHO record ships as a complete release. Client work does the same.
Yes. The vocals are generated; the writing, arrangement, direction, editing and mix decisions are mine. That combination is exactly why the catalog sounds like a catalog and not like a prompt.
I work in a style, an era, a genre and a language. I don't imitate a specific living artist's voice and I don't clone anyone's vocals.
Haitian Creole, French, English and Spanish — with phonetic control over mixed-language lyrics so nothing gets mispronounced.
You choose the tier before work starts, and it is written into the brief:
Standard licence — you use the track wherever we agree, and I may licence it to someone else. Exclusive — nobody else ever gets that track. Full buyout — every right I hold passes to you on final payment. Territory, media and term are set per project. In all three, I keep the right to show the work in my portfolio unless you ask me not to.
Two things I will always tell you straight, because most people don’t. The vocals are AI-generated, and no one can promise in advance that a registrable copyright exists in an AI-assisted recording — that depends on human authorship, which is why the writing, arrangement and direction matter. And I don’t indemnify against third-party claims, because my own tooling doesn’t indemnify me. Anyone who tells you otherwise has not read their own supplier’s terms.
A single track: a first version in about five working days, finished within ten of you approving the direction. An EP of three to five: three to four weeks. A full album: six to eight weeks. The clock starts when I have the brief, the deposit and anything you’re supplying — lyrics, references, brand assets.
Rush work is possible at a premium when the calendar allows. And if notes don’t come back within ten working days, that version is treated as approved, so a project can’t sit half-finished for months.
Two rounds per track, included. A round is one set of notes sent together — and a revision is a change within the direction we agreed: tempo, mix balance, a section length, a vocal take, a lyric line.
A change of direction — different genre, different voice, a fundamentally different song — is a new track rather than a revision. That’s not me being rigid; it’s so the two of us are talking about the same thing when we say “revision”.
No voice cloning of real people. No imitating a specific living artist's voice. No passing AI work off as a named human performer. The AI production is disclosed, here and on every upload.
The more specific you are, the better the first draft — it's the difference between one revision and four. Only your name and email are required.
I'll come back to you at the address you gave me, usually within two business days. If it's urgent, reply to that email and it comes straight to me.