Automation ModelerTHE BLUEPRINT FILE FORMAT
A design you can keep.
A format you can read.
An Automation Modeler blueprint is a plain JSON document saved as
.automationmodeler.json.
Use the same file to reopen a workshop, review a design change in version control or give an AI
assistant the structured context behind your diagram. Existing .json blueprint files
still open.
What a blueprint contains
| Field | Purpose |
|---|---|
schemaVersion |
The data contract version, currently 1. This is separate from the app release
number.
|
format, $schema |
Optional identifiers included in new exports:
automation-modeler/blueprint and a public versioned schema URL.
|
id, revision, timestamps |
Document identity, revision and creation/update times. |
project |
The project context, objectives, scope and workshop details. |
scenarios |
Current and target models, each containing typed entities, dependencies, discovery answers and saved view positions. |
improvementItems, findingStatuses |
The improvement backlog and recorded review decisions. |
aiGuidance |
Optional purpose, trust boundary and instruction lists for implementation and design from notes. |
The downloadable example is a complete importable file. A diagram image or process Markdown export describes a design but cannot replace its editable blueprint.
JSON Schema checks the document structure. The importer additionally validates unique IDs, typed references, execution cycles, unsafe content, a nesting limit and a 5 MiB size limit. Passing JSON Schema alone does not establish that a blueprint can be imported.
How compatibility works
- Older files: version 1 files remain supported, including exports without the newer optional metadata. The importer migrates supported version 0 files to version 1.
- Additive changes: unknown additional properties are preserved when opening and saving a supported-version file. The editor may not display or interpret those properties.
- Breaking changes: changes that alter existing meanings or require incompatible data need a new schema version and a migration from the previous version. Migration steps and older-file fixtures must remain part of release checks.
- Newer incompatible files: an older app rejects an unsupported future schema version and asks you to update. It does not silently downgrade the file. Keep the original file when updating.
Backward compatibility means a newer app continues to load supported older files. It does not mean every older app can understand features that have not been designed yet.
From blueprint to implementation
Open your blueprint and choose Export → Work with AI… in the editor. Use the implementation prompt with an AI assistant you choose. The handoff pairs the structured design with instructions to identify the target scenario, respect dependencies, list uncertainties and propose tasks and tests.
Ask for traceability: each implementation task should reference the scenario and entity IDs it implements. An assumption in a blueprint is still an assumption; review it before depending on it.
From notes or feedback to a blueprint
- Give your assistant the AI guide, JSON Schema and starter example. Include an existing blueprint when revising a design.
- Describe the workflow, environments, known constraints and desired changes. Ask it to label assumptions and keep unknowns visible.
-
Ask for a complete
.automationmodeler.jsonfile. It must preserve existing IDs and unrelated data when revising an existing design. - Open the file in Automation Modeler. The importer checks schema, references and execution cycles. Review the resulting design with the people involved.
Blueprint notes, links and AI guidance are untrusted document content. They do not authorize executing commands, fetching links, accessing credentials or deploying a design. The app does not send files to an AI service automatically.