​

One image API for applications, agents, and your terminal

Generate and edit images, remove backgrounds, upscale, and create reference-based sprite animations through the DreamLayer Agent API, CLI or MCP server. Create logo concepts, product imagery, marketing visuals, print artwork and game assets, then save the results into your project. Ordinary image operations cost one credit each; sprite bundles use frame-count pricing.

​

What DreamLayer decides for you

Routing, model selection, fallback, and execution are handled on our side and stay private. You get opaque job state, sanitized events, and a finished asset. That boundary is deliberate: it is what lets the model behind an operation change without changing your integration.

​

V1 scope

V1 supports four named operations: text-to-image, one-reference-image image-to-image, background removal, and upscale, plus persistent conversations, multiple turns, and durable jobs. It does not support video.

​

Build a reliable integration

Start with the API overview, then use runnable examples, request limits, and safe retries. Connect an AI client using the MCP setup guide, or install the CLI using the command-line guide.

​

What you can build

Install the agent workflows for task instructions, credit checks, recoverable jobs and local Godot/Unity import examples.

• Product image workflows: edit a reference, remove backgrounds, or upscale images while preserving a separate execution for each asset.
• Game asset tools: turn a reference into a sprite-sheet animation with exported frames and playback metadata.
• Agent workflows: connect MCP tools so an assistant can inspect capabilities, check credits, and return a finished asset.
• Scheduled media jobs: use CLI automation or a bounded API queue with saved request identity and recoverable outputs.