AI studio · Paris · Worldwide
We make machines speak.And worlds live.
Paris-based AI studio. Autonomous agents in production for B2B SaaS — and transmedia universes with synthesized voices, 3D and generative narration for artists, labels and brands.
Path 01 · Systems
Build AI systems
Voice AI, autonomous agents, RAG, automation — shipped to production.
Path 02 · Worlds
Build a world
Lore, characters, voice, 3D, generative media — a universe, not a website.
live · webrtc · ~700ms · eu-west-3
Live products
In production. For actual users.
Three systems
Three ways to build with AI.
Not eight competing services — three commercialisable systems. Every technical capability we have lives inside one of them.
AI Operations Engine
Autonomous agents, automation, RAG and analytics — production AI that runs your operations.
- Autonomous agents
- AI automation
- RAG & knowledge
- Data & analytics
- Security & pentesting
- Cloud & scale
- Full-stack platforms
Voice AI Platform
Real-time voice agents in 29+ languages — inbound calls, qualification, support, business integration.
- Inbound & outbound calls
- Lead qualification
- 24/7 support
- CRM & telephony integration
- Post-call extraction
The rare practice · Worldbuilding
Generative World Studio
Canon, voices, images, 3D and interactive experiences — entire transmedia universes, not slide decks.
- Canon & lore
- Character voices
- Generative images
- 3D scenes
- Narrative agents
- Curation pipeline
Worlds we build
AI as a medium for myths.
Beyond agents and SaaS, we build entire universes. Transmedia worldbuilding, generative narrative, immersive 3D — for artists, labels and brands who want a mythology, not a slide deck.
Flagship · in production · expanding
Codemachia — a mythological AI universe.
An AI-native transmedia universe — built in the open. Four novels, four albums, a Yahtzee game against the seven champions, a Champion-matching quiz. The discipline behind it is what we ship for labels, studios and brands.
7
sovereign AIs
2050—2147
narrative arc
3D · Voice · Gen
AI-native pipeline
For artists, labels & brands
Music labels building album universes, brands building mythologies, studios building IPs — clients who need a world, not a website.
Worldbuilding pipeline
Lore architecture, character voices (ElevenLabs), generative visuals (Imagen, Midjourney), 3D scenes (Three.js), narrative agents — orchestrated as one system.
From concept to canon
We don't just illustrate your idea — we build the engine that lets it grow. Lore evolves, characters speak, visuals generate, audiences inhabit.
Selected chapters
Three builds, in depth.
The work behind the studio — a national human-services platform, a living transmedia universe, and an autonomous 8-agent bid pipeline.
Maideo
Home-services SaaS with voice AI recruiting.
Field ops → SaaS platform
Read the chapter
Codemachia
Mythological AI universe — 4 novels, 4 albums, 7 sovereign AIs, playable.
4 novels · 4 albums · 7 champions · playable
Read the chapter
Opportunix
Pay-per-RFP autonomous platform — 8-agent pipeline, ~30 min per bid.
~30 min per bid
Read the chapterMore work
Across domains.
Event tech, civic AI, fintech, prediction games — the same studio solving very different problems.
Friend leagues · 2026
Footfoot
Score-prediction game between friends · 1000 pts/week, top 30% climbs.
Multi-tenant · 2026
The Boom
Multi-tenant PWA for unforgettable private events.
Civic AI · 2026
Democratie
Civic AI assistant for the 2026 Paris elections.
Trading agents · 2025
ikki.finance
Self-learning autonomous quantitative trading engine.
Let's build
Ship your AI project this quarter.
15 minutes. We listen. Then we ship.