Who operates Lians
LIANS AI, CORP., an active New York domestic business corporation, operates lians.ai and is responsible for the Lians commercial services described in this policy. The shortened name LIANS AI may appear in authentication, billing, tax, bank, or payment-provider records. Privacy questions and requests may be sent to privacy@lians.ai.
Information we collect
Account and onboarding data. When you create or use an account, we may receive your name, email address, profile image, authentication identifiers, workspace membership, and the company, role, use case, and tooling information you provide.
Contact and qualification data. If you contact us or request the AI Evidence Readiness Sprint, a demo, or longer implementation work, we collect the information you submit, such as your contact details, company, workflow, deployment environment, technical requirements, referral or campaign information, communications, and any files you choose to provide. We may record whether an opportunity appears to fit our current services. Do not submit passwords, API keys, protected health information, customer content, or other sensitive data unless we have agreed in writing on an appropriate intake method.
Email subscription data. If you previously requested product updates, we may retain your email address, consent record, source page, and campaign attribution until you unsubscribe or request deletion. You can unsubscribe through a message or by emailing privacy@lians.ai.
Billing data. Clerk Billing and Stripe may process existing subscription purchases. Stripe may also process invoices or payment links for the paid sprint and longer implementation services. We and those providers may receive billing contact details, company and address information, tax information, plan or invoice details, and payment status. Payment-card details are handled by the payment provider and are not stored in the Lians application.
Service and device data. We may process request routes, timestamps, status codes, device or browser information, IP addresses used for security and short-lived rate limiting, authentication events, and error reports. Our first-party funnel measurement stores daily aggregate event counts by public page and approved event category. It does not store an analytics cookie, visitor identifier, raw IP address, user agent, email address, form message, or compatibility sample content. We use browser storage where needed for authentication, security, preferences, operation of the service, and session-scoped campaign attribution.
Local compatibility check. The public compatibility check parses a sample inside a same-origin browser worker. The page has no sample-upload endpoint, does not place the sample in local storage or indexed storage, and does not intentionally send its file name, field names, values, or locally calculated hash to Lians. The downloadable mapping report is created on your device and contains contract targets, opaque source-field references, counts, checks, and a local input hash, but no sample values or source field names. Ordinary requests needed to load the page and its static files remain covered by the service-and-device-data description above. Use only synthetic or de-identified samples and do not rely on the local heuristic scan as proof that data is de-identified.
Customer content. In a managed deployment, we process content that a customer sends to the service on that customer's behalf. Application-level per-subject encryption and crypto-erasure apply only where the deployment is configured for those controls and records are correctly associated with a subject. Other storage and encryption protections depend on the selected deployment and its infrastructure.
How we use and share information
We use information to provide and secure the website and services; authenticate users; operate workspaces; evaluate technical fit and route inquiries; send requested product updates; communicate about demos, paid sprints, implementation work, and support; measure aggregate funnel performance; process payments; diagnose errors; prevent abuse; and comply with legal obligations.
Where applicable law requires a legal basis, we process information as needed to provide a service or take requested pre-contract steps; to comply with legal obligations; with consent where we ask for it; and for legitimate interests such as securing, operating, improving, and supporting the service, evaluating business inquiries, and preventing fraud, balanced against the rights of affected people. You may withdraw consent at any time without affecting processing already performed, and you may object to legitimate-interest processing where the law gives you that right.
We do not sell personal information or use managed customer content to train models. We disclose information to service providers that help us perform the functions described in this policy, when a customer directs us to do so, when reasonably necessary to protect the service or comply with law, or as part of a business transaction subject to appropriate safeguards. A qualification result is an internal assessment, not a promise that we will accept an engagement or that a proposed deployment will meet any particular requirement.
Service providers
Depending on the service you use, our providers may include Vercel for website and console hosting; Clerk for authentication and subscription management for the existing Lians console; Cloudflare for authentication-related bot and abuse challenges; jsDelivr for pinned delivery of the Clerk browser library when that asset path is used; Auth0, an Okta service, for OAuth identity, session, and account identifiers used by the hosted OpenAI memory integration; Resend for Auth0 verification and password-recovery transactional email for that integration; Stripe for subscription, invoice, and payment processing; Neon for aggregate analytics, contact, qualification, and email subscription records; Fly.io for hosting the hosted OpenAI memory application and its encrypted PostgreSQL volumes; Fly-provisioned Upstash for Redis for operational state and rate limiting, including that integration's ephemeral rate-limit and cache keys; Sentry for error monitoring; and Google Workspace for business email. Managed deployments may also use infrastructure providers identified in the applicable product documentation, order form, or data-processing terms.
A self-hosted engine does not send workload content to Lians by default. That does not remove the providers used when you visit lians.ai, create a Lians account, contact us, or make a payment. Providers may process information in the United States and other locations subject to their contractual and legal safeguards.
Retention and erasure
We retain account and service information while needed to provide the service and for a reasonable period afterward. Contact and qualification records are retained while an opportunity is active and as reasonably needed for follow-up and business records. Email subscription records remain active until you unsubscribe or request deletion. Daily aggregate funnel counts do not contain a visitor identifier and may be retained for trend analysis. Billing, tax, security, support, and dispute records may be retained for applicable legal, accounting, fraud-prevention, and operational needs. Customer content follows the configuration and written terms for the relevant deployment.
Lians does not receive or retain sample content processed by the local compatibility check. The page releases its working references when a run finishes or you clear the lab, but this is not secure erasure of browser, operating-system, clipboard, download, backup, or device memory. A report you download is stored under your control and should be handled according to your own policies.
Where per-subject keying is enabled and records are correctly associated with the requested subject, crypto-erasure destroys that subject's active content-encryption key and creates a hash-linked proof-of-erasure record. It is not a legal certificate or independent attestation. Backup retention and restore behavior depend on the deployment and backup controls and should be documented in the applicable agreement before regulated or sensitive use. We may retain information that is required by law, needed to establish or defend claims, or already de-identified so it cannot reasonably be linked to a person.
Rights and choices
Depending on where you live and subject to applicable exceptions, you may have rights to access, correct, delete, restrict, object to, or receive a copy of personal information, and to appeal certain decisions. You may also complain to the data-protection or privacy regulator where you live, work, or believe a violation occurred. Send requests to privacy@lians.ai. We may verify your identity and will respond within the period required by applicable law.
If a Lians customer controls information about you in its deployment, direct your request to that customer; where applicable, we process that content on its instructions. Closing a service account, canceling a subscription, and making a privacy deletion request are separate actions. Send an account or personal-information deletion request to privacy@lians.ai; no new account is required to make the request. When a Lians account is deleted, its active workspace state is removed and its namespace keys are revoked, while limited billing, security, dispute, and content-free deletion evidence may be retained where legally or operationally required. You may opt out of non-essential marketing messages using the unsubscribe method in the message or by contacting us.
Security and data roles
We use administrative and technical measures designed to protect information, which may include transport encryption, access controls, monitoring, and deployment-specific at-rest encryption, subject keying, information barriers, and hash-linked audit records. Available controls depend on configuration and do not make any system completely secure. See the security architecture for current product details, and report suspected vulnerabilities to security@lians.ai.
For managed customer content, Lians generally acts as a processor or service provider and the customer determines the purpose and lawful basis for processing. For a self-hosted engine, the customer controls the deployment and its infrastructure. Any data-processing agreement, business associate agreement, residency commitment, or additional security obligation must be agreed in writing before the relevant regulated or sensitive data is submitted.
ChatGPT and Codex hosted memory
Scope and content. This section applies only when you use Lians' hosted memory through the universal OpenAI plugin in ChatGPT or Codex. Lians receives the memory snippets and retrieval queries that you explicitly submit through that integration. It does not apply to a self-hosted Lians engine or describe data handling for other managed deployments.
Identity and client surface. Auth0, an Okta service, processes the OAuth identity, session, and account identifiers needed to authenticate and maintain sessions for this integration. Lians transforms the OAuth account identifier supplied for the integration into an opaque HMAC-based namespace before using it to separate hosted memory. OpenAI, including ChatGPT and Codex, is the client surface and a recipient of the information you submit and the results Lians returns at your request.
Transactional authentication email. Resend sends Auth0 verification and password-recovery transactional email for this integration. For those messages, Resend processes the recipient and sender email addresses, the subject, and the rendered verification or recovery message content, including the applicable link or code, plus provider message identifiers, event timestamps, and delivery, bounce, or failure status. This role is limited to those transactional authentication messages.
Ephemeral Redis data. The hosted integration uses Fly-provisioned Upstash for Redis for ephemeral rate-limit and cache keys. Upstash keeps regular backups outside Fly.io. An expired or deleted active key may remain in those backups until Upstash's scheduled backup-deletion routines remove it; pending deletion, Upstash states that the backup copy is kept beyond use.
Retention and deletion. Active hosted memory is retained for 365 days by default. Expired active content is removed on a scheduled prune cycle, so removal is not promised at the exact instant a memory reaches its retention age. Privacy-minimal audit metadata and HMACs are retained indefinitely unless this policy changes. After you confirm a forget request, Lians crypto-erases the affected active memory. Fly.io encrypted volume snapshots are retained for 5 days; deleted content may remain recoverable from those snapshots until the relevant snapshot expires.
Changes and contact
We may update this policy as our services and practices change. We will post the revised policy, update its effective date, and provide additional notice where appropriate or required. Questions or privacy requests can be sent to privacy@lians.ai or through our contact page.