Privacy policy

What we collect

Account and sign-in data

What you send through the API

Synchronous chat prompts and completions are not stored in the request journal. Asynchronous chat jobs temporarily store the submitted request while it is queued or running and store the answer for retrieval after completion. Uploaded files and media or Tool results have their own job and expiry rules; the journal's metadata-only design is not a promise that all API content is never stored.

For each request we do keep metadata: a request identifier, the workspace and API key it came from, the model you asked for, the model the provider says it served, the provider, the delivery status and HTTP status, prompt and completion token counts, the computed cost and whether that cost is settled, latency, whether a fallback to a second provider occurred, whether the request used a provider key you supplied, and the time. The same request also produces one journal entry per provider attempt, carrying identifiers, token counts and the price version used — no content.

Keys

API keys and management keys are stored only as a SHA-256 hash plus the first characters for display. The secret itself is shown to you once, at creation, and we cannot recover it. If you add your own provider key, it is encrypted with AES-256-GCM under a data key wrapped by an AWS KMS key, and the console only ever shows a hint — the first three and last four characters. Revoking that key erases the stored ciphertext and the wrapped data key in the same operation, so the secret cannot be recovered afterwards, not even by us.

Administrative records

Changes made in the console or through the management API — creating a key, revoking one, changing a spend cap — are written to an audit log recording who acted, what they did, which object it affected and when. Actions taken by BeLive staff on the internal operator plane go to a separate audit log with the staff member’s email address, the action, and the values before and after the change.

Payment data

Where payment is enabled, Stripe receives payment details directly through its embedded card form or hosted checkout, depending on the method. We receive payment references, amounts and payment status, and store saved-card metadata: brand, last four digits, expiry month, expiry year, card fingerprint and issuing country. We retain Stripe customer and payment identifiers, billing names, addresses, contact email and tax identifiers, plus billing, card and IP country evidence and request IP for tax calculation. We never see or store full card numbers.

Technical logs

Our services write structured application logs to Amazon CloudWatch Logs. Those lines never contain prompts, completions, API key secrets or provider keys.

No tracking

The console loads no third-party analytics, advertising or session-recording script. We do not build profiles, and we do not sell personal data.

Google sign-in

If you choose “Continue with Google”, we use Google’s OpenID Connect sign-in. It is handled entirely on our server; no Google token ever reaches your browser from us.

Our use of information received from Google APIs adheres to the Google API Services User Data Policy, including its Limited Use requirements.

You can disconnect AI Hub from your Google account at any time at myaccount.google.com/permissions. That stops future sign-ins; to have the stored data deleted as well, see Deleting your data.

How we use it

We use the data above only to:

• run the service — authenticate you, authorize what your keys may do, route your requests and return the answers;
• show you your own usage, cost and request history in the console;
• calculate what your usage costs, and enforce the spend caps you set;
• send you service email: address verification, password reset, and notices about the service;
• keep the service secure and accountable — abuse and fraud prevention, and the audit logs;
• show a provider, at its request, that use of its models complies with its terms;
• meet legal obligations.

We do not use your prompts, completions or account data to train models. Nothing about your usage is sold or shared for advertising.

Who else sees your data

We also disclose data where the law requires it, and to a successor if the business is transferred — in which case this policy continues to apply until you are told otherwise. There are no other recipients.

Your data is stored and processed in Singapore (ap-southeast-1). Sending a request to a provider may mean the content of that request leaves Singapore, depending on the provider and the model you choose.

How long we keep it

Deleting your data

The console has no self-service account deletion yet. Email us at support@belive.sg from the address on the account and we will delete your account, its organization, its workspaces, its keys and any provider key you supplied, within 30 days. We keep only what we must: payment and accounting records the law requires us to retain, and audit entries needed to show what happened to an account.

You can already do these yourself, without asking us:

• revoke any API key, management key or provider key in the console, immediately;
• correct your name and email address in the console’s settings;
• disconnect Google sign-in at myaccount.google.com/permissions.

Under Singapore’s Personal Data Protection Act you may also ask for a copy of the personal data we hold about you and how we have used it. Write to the address below.

How we protect it

• All traffic to the console and the API is HTTPS.
• The database is encrypted at rest with AWS KMS and is not reachable from the internet.
• Passwords are stored as argon2id hashes, key secrets only as SHA-256 hashes, and provider keys under envelope encryption with an AWS KMS key.
• Service credentials live in AWS Secrets Manager, never in our code or our container images.
• Each organization is isolated: a request carrying one organization’s credential cannot read another’s data, and that isolation is enforced by tests we treat as a release gate.
• Within your organization, the owner, admin and member roles separate who can change billing and members, who can change keys and settings, and who can only use them.
• BeLive staff reach tenant data only through a separate internal plane whose actions are recorded in an audit log. The operator request journal exposes metadata rather than prompt or completion bodies. Async job storage is separate and follows the retention rules above.

No system is perfectly secure. If you believe your account or a key has been compromised, revoke the key in the console and tell us at support@belive.sg.

Cookies and browser storage

There are no advertising, analytics or third-party cookies, so there is nothing to consent to and no banner to dismiss.

Children

AI Hub is for business use and is not intended for anyone under 18. We do not knowingly collect data from children. If you believe a child has created an account, tell us and we will delete it.

Changes

If we change this policy we update the date at the top. For a change that materially affects how we use your data we will also email the address on your account before it takes effect.

Contact

Privacy questions, access requests and deletion requests: support@belive.sg.

BeLive Technology, Singapore.