Krea
35 Intel
Krea is a San Francisco-based AI creative platform offering a unified suite of tools for generating, editing, and upscaling images and video, drawing on multiple underlying AI models through one interface. It was founded by Victor Perez (CEO) and Diego Rodriguez (CTO), who met as students in Barcelona. The company has raised roughly $83 million across seed, Series A, and Series B rounds, reaching a $500 million valuation, with customers including Perplexity, Pixar, LEGO, and Samsung.
Is this you? Sign in with X to claim this profile.
Intel
Systms, a creative studio, describes building its Krea launch demo with Krea Agent in under 48 hours: encoding a creative's process as saved skills, a generation-logging skill, and where the agent fell short. Published by Krea unedited.
Krea Agent is now reachable via MCP, letting agents generate images and video, automate video pipelines, and run creative iteration loops directly inside existing agent workflows.
Krea 2 is now available through an API in Medium and Large variants, differing in size and post-training strength: Medium is cheaper and more consistent on illustration, Large handles photorealism and raw looks. Supports style transfer, moodboard conditioning and a creativity parameter, via Fal, Com
Documents Seedance 2.5's API surface: 30-second single-shot clips, up to 30 image/10 video/10 audio references, and a measured 10+ minute render for a 30-second 1080p generation.
ByteDance's Seedance 2.0 video model is now live on all Krea paid plans, with full-clip lighting and texture coherence and consistent characters across a shot.
Krea releases Krea 3, an image editing model aimed at production use. It claims edits under 8 seconds at $0.06 per image, several times faster and cheaper than Nano Banana Pro, and is distributed through providers like fal, Baseten and Together AI.
Krea released open weights for FLUX.1 Krea, a guidance-distilled checkpoint that slots into the FLUX.1-dev ecosystem, alongside a report on its pre- and post-training. The core argument: optimizing FID, CLIP score and CLIP-derived aesthetic scorers is what produces the recognizable AI look.
Krea engineers detail how they built a custom Virtual Kubelet provider so inference workloads can burst to any available capacity the instant a training run claims their entire shared GPU cluster, including the pod-lifecycle interface they implemented.
Krea Nodes now turns a one-sentence description into a full generation workflow, auto-selecting and wiring nodes into an editable graph instead of manual drag-and-drop setup.
Real fashion-brief comparison of Seedance 2.5 vs 2.0: 2.5 is cheaper and allows longer/more-referenced shots, but 2.0 remains the only one with native 1080p/4K output and better-rated physics and lip sync.
Krea shipped an agent that turns a sentence into a runnable node graph. It reads existing canvas state, proposes a stage plan for approval, repairs type mismatches by inserting conversion nodes, prices the job per node before running, and on edits reprocesses only what sits downstream.