Luminair: Identity, product, interaction, launch
Elion Labs
A living identity and product system for a multi-model AI workspace: one place to think, direct agents, preserve context and move work forward.
Client Luminair · Sector Artificial intelligence, productivity · Scope Identity, product, interaction, launch · Year 2026
01 / Identity
A modular wordmark and a bird built from type-like marks. Black and white hold the system together; green, cyan and blue carry live signal.
02 / Augmented perfectionism
A bird became Luminair’s symbol of agility, speed and resilience, then a living identity system.
Give ambition a living mark.
ChatGPT helped find a symbol with lift, precision and resolve: one bird, already in motion.
Let it take flight.
Higgsfield carried that still into motion, turning a mark into momentum.
Make identity respond.
Codex turned flight into a responsive WebGL effect. Luminair’s mark no longer plays back. It comes alive.
Image generation → motion generation → responsive effect. The sequence remains synchronized so the transformation reads as one idea.
03 / Design system
A precise type system, quiet neutral surfaces and three signal colours. Nothing decorative competes with the work.
04 / The bird, worn
The living mark moves onto simple everyday pieces. Bird only, with its character texture and signal pixels left intact.
Crewneck 01
Cap 02
Tote 03
05 / Product
One workspace for every leading AI, with shared business context and a clear view of work in motion.
AI work is scattered across tools, tabs and lost context.
Luminair brings models, memory and active work into one calm system. Move between specialists without rebuilding the brief. Queue the next request without waiting. Keep progress close from desktop or phone.
Every switch starts with another briefing.
Separate apps keep separate histories, so the same business rules and project decisions get repeated.
Every new thought waits behind the last one.
Without a queue or parallel branches, momentum turns into a chain of blocked turns.
A tiny edit inherits flagship cost.
One-model defaults treat a colour change and a payment migration as if they need equal power.
Closing the laptop hides the work.
Progress, questions and the next request stay tied to the desk instead of following you.
Right model. Same conversation.
Claude, GPT, Gemini and more can work inside one session, each taking the part it does best.
Context that stays useful.
Prices, tone, decisions and rules stay organized once, then reach every model and teammate.
Work moves in parallel.
Line up requests, split independent directions and keep long-running work visible without noise.
Keep going from anywhere.
Check progress, answer a question or queue the next request from phone or tablet.
Same work. Wildly different cost.
Luminair’s routing research shows why one-model defaults fail. Reference cost varies sharply, while routing every small edit through a classifier can add its own delay.
45× spread between lowest and highest models in the same reference token basket.
Reference basket: 4 million input + 1 million output tokens at published api rates, checked 27 september 2026. timing: six synthetic prompts on one mac, not a benchmark. real savings are not yet measured. source ↗
Sessions Direct The Work
06 / In the room
The identity moves into a real technology office: quiet materials, natural daylight and product detail that feels used, not staged.
Brand identity, product design and interaction work created for Luminair by Elion Labs.
Like this project
Posted Oct 13, 2026
A living identity and product system for a multi-model AI workspace: one place to think, direct agents, preserve context and move work forward.
