Save adaptive variable-driven prompts
Repeatable text templates with variables that adapt them at runtime.
Graph creative workflows using over 200 generative and deterministic nodes.
Design using multimodalities across video, image, audio, vector, 3D, code, and text.
















































35+ model providers and counting
Reusable blocks — subgraphs, prompts, variables, databases. Capture a method once, then reuse it everywhere.
Wire typed nodes into a graph, mixing generative and deterministic steps, then run it as one process.
Every run yields a versioned, signed asset with lineage. Browse graphs and assets in one tagged home.
Every node is an intent — Generate Image, Animate, Transcribe — and the model behind it is a swappable implementation. Wire once, swap models without touching a wire. Each node shows its credit estimate up front.
Half of a real pipeline isn't generative — resize, cut a timeline, burn captions, split a clip, extract audio. These run as pure functions: no prompt, no model, the same result every run.
Repeatable text templates with variables that adapt them at runtime.
Bind static values or database-driven ones for consistency across graphs.
Save any graph as a subgraph and drop it into another like a node.
Tables of assets and descriptions, referenced through variables.
Relime's growing library of starter subgraphs will get you going on your creative venture. Create your own subgraphs to simplify your workflows.





























