Project Engine
Text, image & video generation for projects via Model Context Protocol (MCP)
- API Provider Architecture
Flora-style flow built from separate nodes: a Request node (prompt), a Guidance node (image upload), and a Generate node with its own model dropdown — the node adapts to the chosen model: the orange image-input port appears only when the model supports image input, and the output type (image / text / video / 3D) follows the model's capabilities. Wire an output port (●) to a matching input port (purple = text, orange = image); a Generate output can feed another node. Click Next to auto-chain a follow-up scene. Generated images also appear below in the Workflow. Uses the provider keys from "My Model Keys" above.
Load a storyboard, sample scenes (prompts.csv), or our data-pipeline flow (admin).
Enter your GitHub token to save your generated content
Output folder: modelearth/requests/generated
Select GitHub repositories or documentation collections to use as context.