Luna Studio
Music, video and story in one studio. Talk through an edit, hand a rough cut to the assembler, and hear it back.
Open the studio →Where Creativity Begins…
The waveform is the platform's ear. Turn on your microphone and it follows your voice. Nothing is recorded or sent anywhere.
About
SEAMS PRO is the creative platform from SEAMS AI: a place where a song, a script, a cut and a conversation are all the same project, and the tools listen while you work.
Most creative software makes you leave the room to get anything done: export here, upload there, explain the whole idea again to a new tool. SEAMS PRO keeps the thread. Your voice, your files and your collaborators stay in one session, and the platform's agents can see what you see, hear what you say, and act on it.
It runs on infrastructure we operate ourselves: real-time media on our own servers, models we choose and can name, and memory that belongs to your account, not to a vendor.
Purpose
Three commitments shape every product we ship.
Voice is the fastest interface a person has. Our agents hear interruptions, wait when you pause, and stop when you speak.
Share a screen and the agent reads the error, the timeline or the form in front of you, then says what it saw.
Self-hosted media, named open models, and memory kept in your account. Nothing about your work leaves your control.
Products
Each product is a front door to the same platform. Pick the one that matches the work, and the rest is already connected.
Music, video and story in one studio. Talk through an edit, hand a rough cut to the assembler, and hear it back.
Open the studio →The business side of a creative practice: bookings, billing and the webhooks that keep partners in sync.
See Luna One →A voice front desk for hospitality. Guests call, ask and book; the agent answers in their language.
Meet Hotelli →A marketplace with vendors, products and ratings that stay honest, searchable by meaning rather than keyword.
Explore Wonderland →The AI router behind every product: one endpoint, curated open models, and failover that never changes your answer's voice.
Read the router docs →The site reliability engineer you can call. Otto checks the fleet, explains what is wrong, and asks before acting.
Talk to Otto →Docs
Every product speaks the same three calls. Give an agent a persona, a memory and a set of actions, and the voice, vision and telephony come with it.
Create a project, dispatch your first call, and hear the agent introduce itself.
Context, act and checkpoint: what your application answers, and what the platform remembers.
Vision, avatar, noise cancellation, escalation and more, each on, off or per call.
Self-hosted rooms, camera and screen, and what the agent can see on request.
# A dispatch is all it takes: which application answers,
# what it is called, and which features this call runs.
POST /agent/dispatch
{
"agent_name": "voice-01",
"room": "studio-session-4127",
"metadata": {
"backend": "luna-studio",
"agent_name": "Nova",
"features": { "vision": true, "avatar": false }
}
}
# Your application answers three calls, in prose:
POST /api/v1/internal/voice/context # who is calling, what you remember
POST /api/v1/internal/voice/act # do something, say the result
POST /api/v1/internal/voice/checkpoint # the call ended; here is the transcriptAccount
One account across every SEAMS product. Sign in and the agents pick up where you left off.