Publishing and deploying apps | Instabase AI Hub Documentation

Package a project into a versioned app, then productionize it with deployment integrations and API workflows.

Creating an app

This video shows how to publish a project as an app:

Open the project, then select App > Create App.
Set a name and optional description (include purpose and file types when helpful).
Optionally allow others to copy the source project as a template, and choose an app icon for the Hub.

After creation, you can run the app on demand, build ground truth datasets, and run accuracy tests.

Creating deployments

This video covers creating, configuring, running, and monitoring deployments:

Create a deployment with a unique name and description, then select the app, version, and workspace for results and reviewer access.
Configure runtime settings such as secret keys, plus upstream and downstream sources (drives, mailboxes, or custom webhook functions).
Set email or webhook notifications for deployment events, then run and monitor activity from the deployment view.

Using deployment integrations

Integrations let you pull data from upstream sources or push data to downstream sources. Before you configure integrations, ensure your target data source or destination is configured under the workspace’s Data tab. Supported data connections include S3 buckets, Google Cloud Storage, Azure Blob Storage, Gmail, and Outlook.

Add integrations from the deployment’s Configuration tab:

Upstream — Automatically pull input files from a connected drive or mailbox for processing.
Downstream — Send processed results by email, write them to a connected drive, or use a custom function to deliver the results to other systems.

Using AI Hub APIs

AI Hub provides a REST API that allows you to drive automation programmatically instead of interacting with the UI manually. With the API, you can run deployed apps, manage files and batches, and integrate document processing into your own systems and workflows.

If you’re not confident working with API calls and integrations, this lesson is optional. The core functionality required to work with deployed apps is supported in the AI Hub user interface.

Before you begin

Read through the Using the API and SDK guide.
Familiarize yourself with how API authorization works in AI Hub, and follow the steps to get your own API token.
Review the endpoint reference for the Batches, Runs, Files, and Secrets endpoints.

This video demonstrates an end-to-end use case: using the API to upload files, running the app with the files as input, and retrieving the results.

Create a batch and upload files to the batch.
Run the app, with the batch as your input.
When the run is complete, retrieve the processed results.