Privacy Policy

Effective October 20, 2026

This policy explains what Velobyte, Inc. (d/b/a InferOptimal) collects when you use inferoptimal.com, the console, dedicated endpoints, the console API, the MCP server and the command-line tools, and what we do with it.

What we collect

• Account details: your email address, name and the ID from the sign-in provider you use (Google, GitHub or an email link).
• Billing details: your credit balance, purchases and charges. Stripe collects card details; we never see or store card numbers.
• Endpoint records: the models, hardware, settings and status of your endpoints, and credentials you give us to connect a cloud account or a private model repository.
• Usage data: numerical measurements such as token counts, request timings, throughput, errors and GPU utilization.
• Your content: prompts, files and outputs pass through our gateway on their way to your endpoint. Our analytics keep numbers about requests, not their text. Startup and failure diagnostics, and custom containers you run, can keep more; the data handling page lists each store and how long it lasts.
• Messages: what you send us by email or through support.

How we use it

To run, bill for, secure and support the service; to send you account, billing and endpoint emails (for example, when an endpoint is live or idle); to improve serving settings from usage data; and to meet legal obligations. We do not use your content to train models, and we do not sell personal data. We do not run advertising or analytics trackers on these pages; the console sets only the cookies sign-in needs.

Who we share it with

Service providers that process data for us under contract: Cloudflare (network, hosting and storage), the GPU provider that hosts your endpoint (such as Modal, RunPod, Verda, Lambda, Nebius or SF Compute), Supabase (sign-in), Stripe (payments), Resend (email), Google Fonts (console typefaces) and Hugging Face (model downloads). If startup inference is on for a launch (it is on by default for GLM-5.3-Flash and shown before you launch), requests made while your GPUs start go to the model maker's API, for GLM that is Z.ai. Turn it off and they wait for your GPUs instead. Endpoints in your own cloud account run on your provider. We also disclose data when the law requires it, or as part of a merger or sale of the business with notice to you.

How long we keep it

Account and billing records are kept while your account is open and afterwards for as long as tax and accounting law requires. Endpoint metrics and diagnostics expire on the schedules on the data handling page. Saved cloud and repository credentials expire after 30 days or when you disconnect them.

Your choices and rights

You can ask us to access, correct, export or delete your personal data, or to close your account, by emailing team@velobyte.ai. Depending on where you live, you may have further rights, including to complain to a data protection authority. We will answer within 30 days.

Security

Traffic is encrypted in transit. Platform and endpoint keys are shown once; we keep hashes of them to check requests. Access to customer records is limited to the people who run the service.

Changes

We will post changes here and email account owners before a material change takes effect.

Contact

Velobyte, Inc. (d/b/a InferOptimal), team@velobyte.ai.