Pricing

A PySDK local runtime is any supported runtime that lets you run models locally on accelerators or vendor runtimes, beyond the default CPU path. PySDK local runtimes are used when you want higher performance, lower latency, or on-device processing.

PySDK local runtimes include support for runtimes such as HAILORT, DEEPX, OPENVINO, AXELERA, MEMRYX, AKIDA, and RKNN.

Professional includes 1 user and Enterprise includes 5 users. You can add additional users for an extra monthly charge:

• Professional: $5.99 per additional user, per month
• Enterprise: $4.99 per additional user, per month

Additional users are billed based on the maximum number of additional users on your Workspace at any point during the billing period. If you add users and remove them later in the same period, charges are based on the highest additional-user count during that period.

Professional includes local device licenses for up to 3 devices, and Enterprise includes up to 10 devices. If you need more devices, additional devices are billed monthly:

• Professional: $4.99 per additional device, per month
• Enterprise: $1.99 per additional device, per month

Devices are billed based on unique devices per billing period. During the billing period, we count the unique devices that request or renew a license at least once. If a device requests or renews a license at any point in the billing period, it is included in that period’s device count and charge.

Yes. You can upgrade as your needs grow and switch plans as your team or usage changes. If you downgrade, your included limits and features adjust to match the new plan at the start of the next billing period.

If you cancel a paid plan, your Workspace stays on the paid plan until the end of the current billing period. You keep access to paid-plan features through that date, and usage-based features continue to work as normal.

On your normal billing date, you will receive your final charge for the paid plan plus any usage-based fees from that billing period. After that date, your Workspace automatically switches to the Free plan and there are no further paid-plan charges.

No. Cancellations and downgrades take effect at the end of the current billing period. This avoids mid-cycle proration and keeps billing predictable. You keep paid-plan access through the end of the period.

An AI Hub inference is one model run processed by AI Hub. In most cases, that maps to one API request for one input item (e.g., one image).

When you reach your monthly limit, you can upgrade to a higher plan for more included inferences. Inferences will pause until the next billing cycle or until the plan is upgraded.

Cloud Compiler is billed per successful compile at $4.99 on Professional and Enterprise. A successful compile is a Cloud Compiler task that:

• finishes with a Success status
• produces a valid compiled output artifact for the selected target
• is available to download or use (e.g., to run on a supported device)

You are not billed when a compile fails due to errors, times out, is canceled, or does not produce a successfully compiled output.

Application Packages are pip-installable packages that provide a simple, PySDK-style API for building complete workflows. They are available with the Professional and Enterprise plans with separate additional pricing. You can install them with pip and follow our documentation to get started quickly.