Game asset creation
Characters, worlds, props, and interface pieces for 2D games, held to one visual direction as you iterate.
See game asset creationAbout DreamLayer AI
What DreamLayer AI is.
DreamLayer AI is an image generation and editing platform. You describe the image you need, add a reference when identity or style should carry forward, and refine the result over several turns. It runs in the browser at dreamlayer.io and through an API, a CLI, and an MCP server for products and coding agents.
DreamLayer AI also publishes DreamLayer Eval, an open-source project for benchmarking image and video generation models. The two are separate: the product makes images, the research project measures models.
What you can do with it
Three workflows, one image platform.
Each page below describes one workflow in detail, with real outputs and its own questions answered.
Photo culling and editing
Review a full shoot, group similar frames, keep the strongest moments, and edit the set in one workflow.
See the photography workflowImages for products and agents
Text-to-image, image-to-image with one reference, background removal, and 2x upscale through one managed API, CLI, or MCP server.
See the Agent APIHow you use it
Four interfaces to the same account, images, and credits.
- BrowserDescribe, attach a reference, and refine on the canvas at dreamlayer.io.
- APIGenerate and edit images from your own product through the DreamLayer Agent API.
- CLIRun the same operations over local files from your terminal with the dreamlayer command line tool.
- MCP serverGive Claude Code, Codex, Cursor, and other MCP clients image tools through the DreamLayer MCP server.
Pricing
One credit per finished image.
No subscription and no seat fees. Credits do not expire, and an unfinished request restores the reserved credit. Details are on the pricing page.
Open-source research
DreamLayer Eval is our benchmarking project, not the product.
DreamLayer Eval benchmarks image and video diffusion models with fixed prompts, seeds, and configs, and logs metrics such as CLIP Score and FID so comparisons can be reproduced. It is released under the GPL-3.0 license and is documented on the DreamLayer Eval page, with the code on GitHub and results on the model leaderboard.
If you read that DreamLayer AI is a benchmarking platform, that description belongs to DreamLayer Eval. The DreamLayer AI product does not require it and runs as a managed service.
Contact
DreamLayer AI is in beta.
Questions, workflow requests, and issues go to mackenzie@dreamlayer.io.
Start here
Make your first image.
Open the app in the browser, or get an API key and use the interface that fits your work.