Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.iblueprint.ai/llms.txt

Use this file to discover all available pages before exploring further.

iBlueprint lets you build AI workflows visually by connecting nodes in a linear chain called a Blueprint. Each Blueprint is a self-contained, runnable package — it holds its own nodes, variables, version history, and access settings — so you can build, test, share, and publish workflows without touching any infrastructure.

​
What is a Blueprint?

A Blueprint is a sequential chain of nodes where the output of each node flows automatically into the input of the next. You define the logic once, then run it on demand from the web app, the API, or the CLI. Blueprints support dynamic data through a variable system, letting you pass different inputs at execution time without editing the workflow itself.

​
Key concepts

Nodes

Nodes are the individual steps inside a Blueprint. Each node has a type (such as prompt, api, or conditional), a configuration (the settings specific to that type), and an output that is passed downstream. iBlueprint supports more than twenty node types covering LLM calls, image generation, code execution, HTTP requests, human approvals, and more.
Nodes execute one after another in the order you arrange them on the canvas. There are no parallel branches — each node receives the full output of its predecessor and produces its own output before the next node starts. This makes the data flow predictable and easy to reason about.
Variables let you inject dynamic values into any node configuration using {{variable_name}} placeholders. You can define variables at the Blueprint level and then supply concrete values each time you run the Blueprint. This keeps your workflow reusable across different inputs.
Every change you save can be recorded as a new semantic version (major, minor, or patch). iBlueprint maintains a full snapshot history so you can review what changed between versions. You can also fork any Blueprint you have access to, creating an independent copy that starts at version 1.0.0.

​
Blueprint lifecycle

Blueprints follow a simple two-stage lifecycle:
1

Draft

When you create a Blueprint it starts in draft status. You can edit nodes, update configurations, run test executions, and iterate freely. Drafts are private by default — only you and your collaborators can see them.
2

Published

Once your Blueprint is ready, you publish it. A published Blueprint can be shared with your organization or listed on the public Marketplace for other users to discover and fork. Publishing does not prevent further edits — you continue saving new versions while the Blueprint remains published.
You can unpublish a Blueprint at any time, which returns it to private draft status without deleting any version history.

​
Explore the Blueprints section

Node types

Browse every node type — LLM, image, data, logic, integration, and human nodes — and learn when to use each one.

Variables

Learn how to define Blueprint variables and pass dynamic values at execution time using {{variable_name}} syntax.

Execution

Understand how execution flows node-by-node, how to trigger a run from the UI or API, and how to monitor progress.

Versioning

Save new versions with semver bumps, review history, fork Blueprints, and publish to the Marketplace.