If you've used Layer before, welcome to the new agentic app — it works differently from the version you knew. The short version: instead of picking a tool and driving every step yourself, you now describe what you want and a creative Agent does the heavy lifting alongside you. Here's what changed, why, and how your old habits map across.
💡 The legacy app is still available — for now — at legacy.app.layer.ai. We're keeping it around during the transition, so you can head back if you need to. But the new app is where Layer is going, so it's worth getting comfortable here.
What the old Layer was
Legacy Layer was a powerful creative toolbox. From the home screen you picked the tool for the job — 2D, 3D, Video, Refine, Edit, Realtime, Canvas, or Workflows — then chose a model, set your parameters, and generated. It gave you fine-grained control, but you drove every step and moved between separate tools.
What the new Layer is
The new app is agent-first. You start by describing what you want in plain language, and the Agent works out the how — choosing the model, writing the detailed prompt, generating, and refining with you. It all happens in one place: a conversation paired with a visual board.
What changed, and why
You talk to an Agent, not a control panel
Before: pick a model, tune parameters, craft the perfect prompt.
Now: describe the outcome — the Agent handles model choice and prompt craft (you can still get specific when you want control).
Why: it lowers the barrier and lets you focus on creative intent instead of settings.
Everything lives on one canvas
Before: separate tools and screens for generating, editing, refining, and arranging.
Now: a single session — chat on one side, a board (canvas) on the other — where you generate, iterate, compare, and organize without switching tools.
Why: less context-switching; your conversation and your results stay together.
It remembers what you've done
Before: each generation started from scratch.
Now: the Agent keeps the context of your session — plus your workspace and project settings — so it builds on what came before instead of making you re-explain.
Why: continuity across a whole piece of work.
Projects keep work together
Before: assets piled up in your Drive.
Now: group related chats and assets into Projects for a character, a campaign, or a deliverable. (Everything you make still lives in the Library — the evolution of Drive.)
Why: real work spans many generations, and Projects give it a home.
Built for game marketing and UA at scale
The new app leans into the creative game teams actually ship — especially user-acquisition (UA) and marketing creative: many variations, formats, and iterations, produced fast. The Agent, the canvas, and reference sets are designed to make that volume manageable.
Why: it's where Layer is focused, and where the agentic approach pays off most.
How your old habits map across
In legacy Layer you… | In the new app you… |
Picked a tool tile (2D / 3D / Video / Refine) | Just describe what you want — the Agent routes to the right type |
Chose a model and set parameters | Let the Agent pick (or tell it your preference) |
Wrote a detailed prompt | Describe the outcome; the Agent enriches it |
Used Refine / Edit tools | Ask in chat, or use Transform on a result |
Composed in Realtime / Canvas | Work directly on the session board |
Browsed assets in Drive | Browse the Library; group work in Projects |
Built a Workflow | Workflows are still here for repeatable, multi-step pipelines |
Trained a custom style | Use style training and reference sets to stay on-model |
Getting comfortable
The fastest way to settle in is to make something.
What's next
Meet the building blocks → [What is Layer?]
Make your first thing → [Generate your first asset]
Find your way around → [A tour of your workspace]



