Terms & Conditions

If you request a refund, a non-refundable processing fee of 5% of the original payment amount will be deducted from the refunded sum. This fee covers payment processing charges levied by our payment provider (Stripe), which are retained by the provider even when a transaction is refunded and are therefore not recoverable by FlashAlpha. The remaining balance will be returned to your original payment method.

We genuinely want you to push the platform. The Alpha plan is built for serious strategy work, and we actively support power users running ambitious research, custom logic, and high-throughput workloads. We see that as the platform working as intended, not as a problem.

The Alpha plan runs on shared compute nodes, meaning the same resources are used by other Alpha members at the same time. To keep the experience fast and reliable for everyone, we ask all users to apply reasonable judgment. Sustained, disproportionate, or runaway consumption (for example, malfunctioning or unbounded scripts, tight retry loops, or workloads that materially degrade performance for other users on the same node) goes beyond what a shared environment can fairly absorb.

We are a developer-friendly platform and we want your code to run well here. If a script is consuming a disproportionate share of resources, our first step is usually collaborative: we will point out the bottleneck and ask that the script be reviewed and optimised, since most heavy load on a shared node comes from inefficient or unbounded logic rather than legitimate scale. To keep shared compute healthy for everyone, we expect workloads to follow reasonable engineering standards, including bounded loops and retries, sensible polling intervals and back-off, cleaning up idle or runaway processes, and not relying on the platform to absorb code that has not been profiled or tested. We are happy to help identify what to improve.

If usage reaches that point, our first step is always a conversation, not a penalty. Where optimisation is not sufficient and the workload is simply large by nature, the right path is more room to work: we will recommend upgrading to Alpha Pro with dedicated compute, where you get isolated resources sized for demanding and continuous workloads and no shared-node constraints. Dedicated compute is the supported home for heavy, long-running, or experimental strategies, and it lets you scale without affecting, or being affected by, other users.

Where usage on a shared node is, in our reasonable assessment, excessive, disruptive, or detrimental to other users, and is not resolved through an upgrade to dedicated compute or other agreed remediation, FlashAlpha reserves the right to throttle, suspend, or revoke access to the affected service to protect overall platform stability. We will always aim to give notice and offer a path forward before taking such action, except where immediate intervention is necessary to preserve service integrity for other users.