> For the complete documentation index, see [llms.txt](https://docs.unitlab.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.unitlab.ai/solutions/architecture-and-adoption-guides/platform-architecture.md).

# Platform architecture

How data moves through Unitlab: Data Assets, projects, workflows, quality checks and releases.

Every Unitlab use case follows the same path, from raw files to a versioned release.

<figure><img src="https://2337177515-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi6jXAxtoN6UXTLAJFexq%2Fuploads%2FdJBbf1bIZ7MBbvYh3HX1%2Fsol-workflow-graph.webp?alt=media" alt="Workflow graph: Project, Annotate, Review and Complete stages, with a rejected route back to Annotate" width="800"><figcaption><p>A project's workflow: items move from Annotate to Review to Complete, and rejected work goes back.</p></figcaption></figure>

| Layer           | What it holds                                                                    | You use it to                                         |
| --------------- | -------------------------------------------------------------------------------- | ----------------------------------------------------- |
| **Data Assets** | Every uploaded or cloud-linked file, folders, Data Groups, datasets and versions | Store and curate data once, reuse it in many projects |
| **Project**     | Attached data, ontology, Instructions, members                                   | Define what to label and how                          |
| **Workflow**    | Stages: Model, Annotate, Review, Agent, Webhook, Quality Gate, Complete          | Route each item through people and models             |
| **Workbench**   | Editor for every data type, with multi-tile layouts                              | Label and review                                      |
| **QA**          | Quality checks, consensus, benchmarks, issues                                    | Find and fix problems                                 |
| **Releases**    | Versioned exports in COCO, YOLO, UUEF, GeoJSON and more                          | Deliver data to training                              |

Access is controlled by workspace roles and permissions, and changes are recorded in the audit log. Developers can drive the same flow through the SDK, the API or the MCP server.
