What it does
Scenario gives teams access to more than 650 image, video, 3D and audio models from over 40 providers in a single browser workspace. You can compare models on the same brief, keep a look consistent using references, and train a custom model on your own artwork when a project needs a style that references cannot capture. Repeatable work is built as chains of nodes on a canvas and saved as workflows or small apps that teammates run from a simple form, with the cost shown before each run. A Node Agent turns a written brief into a node graph, and an assistant called Genie chats with your workspace to pick models and fill in forms. The same models, workflows and assets can be reached from Claude Code, Codex or Cursor through Scenario MCP, or from code through a REST API with TypeScript and Python SDKs. An experimental Blender plugin brings generation into the viewport. For enterprises, the page describes approved model lists, per-project spend caps, an optional IP Detection add-on, single sign-on, audit logs and SOC 2 Type II compliance. It says customers own what they generate and that their content does not train models without written consent.
Capabilities
- 650+ image, video, 3D and audio models from 40+ providers
- Node-based workflows that can be published as apps for teammates
- Custom models trained on your own reference art
- Node Agent and Genie assistant for building workflows from a description
- MCP connection for Claude Code, Codex and Cursor, and a REST API with SDKs
- Enterprise controls: approved models, spend caps, SSO and audit log
When to use it
- Your team produces images, video or 3D assets at volume and needs a consistent look
- You want to compare several generative models on one brief
- You need to call generation models or saved workflows from your own pipeline or coding agent
- Your organisation needs spend caps and approved-model rules for generative AI use
