Krea capabilities

Krea exposes a unified REST API for programmatic image generation, video generation, and upscaling, plus a no-code node editor for chaining models into visual workflows that can themselves be deployed as API endpoints. This is the portability path for moving Krea generation into an external pipeline or another product.

An AI-powered upscaling and refinement tool for images and videos that increases resolution, recovers detail, and improves clarity. It works on both newly generated Krea assets and uploaded images, with an optional prompt that guides how the model fills in missing detail (Krea's guidance: describe the final image you want, not the change you want made).

Krea is a model-aggregator creative suite: rather than a single proprietary model, it exposes 64+ image, video, and 3D models in one interface, including third-party models (Flux, Ideogram, Gemini, Nano Banana, Runway, Luma) and Krea's own first-party Krea 2 model. The pitch is access to the latest models on release day from a single subscription.

Krea meters usage in compute units consumed per generation and offers four individual plans (Free, Basic, Pro, Max) plus two team plans (Business, Enterprise). Yearly billing saves 40% on individual plans and 20% on Business. Higher tiers raise the monthly compute allocation and unlock higher upscale ceilings, more LoRA training, and unlimited concurrency.

Krea's signature feature: a live generation surface where the canvas is the input and AI output updates instantly as the user draws, types, or feeds a camera/screen, with no render button or queue. Krea positions itself as the market leader in realtime image generation, turning simple primitives into photorealistic images in under ~50ms.