Privacy notice.
How SynchroX handles information submitted through this website. Updated November 2, 2026.
Customer sign-in.
Customer sign-in uses Supabase Auth with Google, GitHub or Discord. Your chosen provider shares basic identity and email information with Supabase and SynchroX to establish your account. SynchroX uses the verified account ID to separate fleet records. No mail, calendar, Drive, repository, bot or server access is requested. Authentication tokens are stored in secure, HTTP-only cookies on this browser, not browser local storage. Your provider handles its own account credentials. Existing ChatGPT demo accounts are separate.
Automatic sign-out.
SynchroX uses an essential browser-session cookie and short-lived tab-presence records to end access when you leave. When the last tab reports an exit, a 10-second grace period allows a refresh or quick return. If the browser cannot report the exit, server access expires three minutes after its last heartbeat. Sleeping, suspended or offline browsers may need to sign in again. Switching tabs or minimizing alone does not request sign-out. This ends this browser’s SynchroX access, not your provider account or another device’s session. Presence records contain an account ID, a hashed session identifier, random document identifiers and timestamps; they do not record browsing history.
Your research workspace.
Your existing customer account, sign-in methods and recovery remain in place. Google, GitHub, Discord and username/password authentication use Supabase. Username/password accounts retain a verified email for authentication and recovery; account emails are sent through Resend. Robot metadata and experiment records you submit are stored with your immutable account ID in the website database and are private to your account. Each run retains its source reference, observation timestamp and server-saved time. Registration does not discover or control hardware. Do not submit patient data or clinical records. Use the privacy contact form for correction or deletion requests.
Previous server workspace.
The previous server console uses Supabase-managed customer sign-in with Google, GitHub and Discord and an immutable account ID to separate workspaces. Your fleet starts empty. When you create a collector credential and run the collector with sending enabled, it submits NVIDIA-reported UUID, model, serial when available, PCI address, driver version, memory, temperature, utilization, power, timestamps, and query status. Windows AMD mode instead submits model, driver version and the Windows PnP device-instance identifier, with unavailable metrics left empty. That host-local identifier is not a unique GPU serial. System mode also submits CPU model and core topology, RAM capacity and available Windows module part numbers, disk models and capacities, reported serials or host-local locators, and per-category availability. Linux RAM is OS-usable total; physical DIMMs are unavailable in this collector. Your chosen machine labels, approved maintenance plans, verification evidence and notes are stored with your account in the website database. No process lists, host names, usernames, or files are collected by the collector. The server stores credential hashes, not recoverable tokens. Revoking a credential stops future reports; historical records remain. The console retains the latest 100 inventory and connection events per machine and displays the latest 200 maintenance notes. Use the privacy contact form to request deletion.
Information you provide.
Contact and interest forms collect your name, email, request type, and message, along with any organization or role you choose to provide. We record the submission time and the version of the contact consent you gave. Please do not include passwords, tokens, or confidential infrastructure details.
How we use it.
SynchroX uses these details to review and respond to your inquiry, discuss a potential pilot or partnership, or record your waitlist interest. Sending a form does not create a paid account or guarantee a pilot place. No automatic confirmation email is currently sent.
Storage and access.
Submissions are stored in the website database through ChatGPT Sites and Cloudflare. They are not posted publicly. The inbox and export are restricted to the Site owner. Hosting providers may process technical request information to operate and protect the site. The form uses a short-lived hashed network address for request limits.
Demo and product guide.
The mini demo and preset-answer product guide run in your browser. Their interactions are not sent to the inquiry inbox. Scenario results and chat messages reset when the page reloads. If you enable optional preferences, your selected demo rack can be remembered on this browser. The public robotics demo requires no sign-in and keeps simulated robots and experiment records in browser memory. Its countdown uses session storage; it never adds hardware to a real account.
Cookies and browser storage.
Essential storage supports privacy choices, sign-in, and security. Optional preference storage is off until you enable it. This website does not install analytics or advertising tools. You can reopen Cookie settings in the footer at any time, including to withdraw permission for optional storage. Read the Cookie notice for the current categories and storage details.
Questions and requests.
Use the contact form and select Privacy request to ask about, correct, or request deletion of information you submitted. Include the relevant email address and a clear description of the request. The team may need to verify your identity before acting. The public contact email and business address are currently placeholders.