Privacy Policy
Last updated: November 5, 2026
Previous versionsSeptember 17, 2026September 30, 2026
1. Scope and our roles
This Privacy Policy explains how we handle personal data when you create an AMP account, use the AMP console, top up your balance, or send requests through the gateway. It does not cover the open-source Subquill desktop application, which runs locally with no backend of its own, stores your API key and optional Bilibili cookie in local files, and sends us no telemetry.
We act in two different roles. For account, billing, support and usage data we are an independent controller: we decide why and how that data is used. For personal data contained in the content of an API request, we act only as a processor (or subprocessor) on your instructions and on the terms of the Data Processing Terms. Keeping these roles separate is why request content is treated differently from account data throughout this policy.
The controller of the data described in this policy is surfacecyber, a Delaware corporation. You can reach us at [email protected]. Our registered agent for service of process is Legalinc, 131 Continental Drive, Suite 305, Newark, DE 19713, USA; this is a registered agent address, not our place of business.
2. Data we collect
| Category | Examples | Why we use it |
|---|---|---|
| Account data | Email address and display name supplied by Google or GitHub, OAuth provider identifier, account creation date | To create and secure your account and sign you in |
| Billing data | Prepaid balance, top-up amount and date, Stripe transaction reference, chargeback status | To credit your balance, keep accounting records and handle disputes |
| Usage metadata | Model called, input and output token counts, request timestamp, response status, key identifier | To meter your balance, show usage in the console, enforce rate limits and detect abuse |
| API key records | Key name, prefix, creation and last-used date, revocation status | To let you manage credentials and investigate leaked keys |
| Support correspondence | Messages you send to [email protected] | To answer your question and keep a record of the issue |
| Request content | Prompts, transcript chunks and completions passing through the gateway | Only to forward the request to the provider you selected and stream the response back |
3. API request content
AMP is a pass-through gateway. Your request is forwarded to the model provider behind the model you selected and the response is streamed straight back to you. Prompts and completions are not written to our databases or logs and are not retained after the response is delivered; they exist only in memory for the time needed to relay the request.
- We do not use prompts or completions to train models, and we do not sell or share them for advertising or profiling.
- Only usage metadata (model, token counts, timestamps, status) is logged, to meter your balance and protect the Service.
- Upstream providers process request content under their own terms and privacy policies, which may include short-term abuse monitoring. We do not control those practices and do not represent that every provider offers zero retention.
- Please do not include sensitive personal data, credentials, or third-party confidential information in prompts. If you need stronger isolation, point the desktop app at your own OpenAI-compatible endpoint.
4. Why we are allowed to process it
- To perform our contract with you: creating your account, routing requests, metering usage and crediting your balance.
- To comply with legal obligations: keeping accounting and tax records, and responding to valid legal requests.
- For our legitimate interests: preventing fraud and abuse, securing the Service, keeping backups, and understanding aggregate demand so we can plan capacity. We do this using usage metadata rather than request content.
- With your consent where required, for example for optional communications. You can withdraw consent at any time without affecting earlier processing.
5. How long we keep it
- Request content is not retained after the response is delivered.
- Account data and API key records are kept while your Account exists and are removed when you delete it, except where a specific legal claim requires us to keep a record.
- Usage metadata is kept for as long as needed to bill accurately, resolve disputes about charges and detect abuse, after which it is deleted or irreversibly aggregated.
- Billing and accounting records are kept for as long as applicable accounting and tax rules require, regardless of account deletion.
- Routine backups are cycled out on a short schedule; data in a backup is not restored to active use except for disaster recovery.
6. Third parties and subprocessors
We use a small number of third parties to run the Service. The list below is also the list of subprocessors referred to in the Data Processing Terms; we update this page before a new subprocessor begins processing personal data.
| Third party | What it does | Data it may see |
|---|---|---|
| Google and GitHub | Sign-in and identity verification | Your account email, display name and provider identifier |
| Stripe | Processing card top-ups | Payment and transaction details; we never receive full card numbers |
| Model providers offered in the console | Generating the Output for the model you select | The request content you send to that model, under its own terms |
| Our hosting provider | Running the gateway and console on a virtual private server | Account, billing and usage data; request content in transit |
| Cloudflare | TLS termination and network protection for our sites | Network metadata such as IP address and request headers |
We require subprocessors to be bound by confidentiality and data-protection obligations, and we remain responsible to you for their performance. We do not sell personal data and do not share it for cross-context behavioural advertising.
7. Security
We protect data with measures appropriate to a small hosted gateway: TLS in transit, encrypted storage for account data, role-based and least-privilege administrative access, API keys that are never shown again in full after creation, logging limited to usage metadata, regular backups, and prompt revocation of credentials when someone leaves the team. We review access when the team or the infrastructure changes.
8. Personal data breaches
If we become aware of a personal data breach affecting your data, we will investigate it, take reasonable steps to contain and remediate it, and notify you without undue delay where the law requires or where the breach is likely to affect you. Where you are a controller, we will give you the information and cooperation you reasonably need to meet your own notification duties. Notification is not an admission of fault.
9. Your rights and choices
Depending on where you live, you may have the right to access, correct, export or delete your personal data, to object to or restrict certain processing, to withdraw consent, and to complain to your local data-protection authority. We will not discriminate against you for exercising these rights.
- You can view your balance and usage in the console, revoke API keys there, and delete your Account there, which removes your profile and key records.
- For access, correction, export or anything else, contact us at [email protected]. We may need to verify your identity first.
- We will respond within the period required by applicable law. If we cannot comply with a request, we will explain why.
- Accounting records that we must keep by law are the main exception to deletion; we will tell you when that applies.
- We do not sell or share personal data as those terms are defined under California law, and we have no advertising trackers on the console.
10. International transfers
Our infrastructure, our subprocessors and the model providers you choose may be located in countries other than yours, so your data may be transferred across borders. Where a transfer of personal data out of the EEA, the United Kingdom or Switzerland requires a legal safeguard, we will put an appropriate mechanism in place with you, such as the applicable standard contractual clauses, before making that transfer. Email [email protected] if you need details of the mechanism that applies to your Account.
We are not certified under the EU-U.S. Data Privacy Framework and do not rely on it. Transfers of request content to a model provider are also governed by that provider's own terms and transfer mechanisms.
11. Children
The Service is not directed at children and we do not knowingly collect personal data from them. If you believe a child has created an Account, contact us and we will delete it.
12. Changes and contact
We may update this policy as the Service changes. Material changes will be announced in the console at least 3 business days before they take effect; the current version is shown on this page and earlier versions remain available here.
For privacy questions, data requests or complaints, email [email protected]. You can also raise a question about an Upstream provider directly with that provider.