Aplum AI Privacy Policy
Effective date: July 25, 2026 Version: 2026-07-25-v4
This Privacy Policy explains how Aplum AI LLC ("Aplum AI," "we," "us," or "our") collects, uses, discloses, and retains personal information when a person visits, creates an account for, purchases, or uses Aplum AI. It also explains the choices available to users. Contact founder@aplumlabs.com with privacy questions or requests.
1. Scope and roles
This Policy covers Aplum AI's public website, customer account, model-evaluation application, billing and support functions, and related transactional communications (collectively, the "Service"). The Service may be offered to United States businesses and adults acting for them ("Business Customers") and, only when Aplum AI's separate individual-sales launch controls are enabled, to adult United States residents purchasing and using it personally ("Individual Customers"). It is not directed to children.
For account administration, security, billing, direct support, and Aplum AI's own business operations, Aplum AI determines why and how personal information is processed. For prompts, test materials, model inputs and outputs, reports, and similar material submitted through an account ("Customer Content"), the Customer decides what to submit and which AI providers and workflows to use. A Business Customer determines the business purpose and is responsible for required notices and authority for personal information it submits. An Individual Customer submits Customer Content for that person's own purposes and is responsible for having rights to submit any information about another person.
This Policy does not govern an AI provider, payment service, linked website, or other third party acting under its own terms and privacy notice. Their practices may differ from ours.
2. Information we process
Depending on how the Service is used, we process the following categories.
- Account and eligibility information: email address; password hash; selected Business Customer or Individual Customer type; business name and role when applicable; country; age, US-location, business-use and authority or individual-use, and restricted-use attestations; email-verification status; account creation and update times; plan and access status; and administrator status for Aplum AI personnel.
- Availability-interest information: name, email address, and submission time when a visitor asks to hear about Aplum while new-account registration is unavailable.
- Contract and consent records: the Terms and Privacy Policy versions presented; acceptance or acknowledgment method and time; document hashes; selected subscription plan, displayed price, billing cadence, and recurring-charge authorization version; and, if Research Mode is enabled and separately accepted, research authorization, withdrawal, disclosure version, plan, source, and timestamps.
- Customer Content: test batteries and reviewed imported rows; prompts, rules, scenarios, personas, narrative tests, chat messages, model inputs and outputs, grades, annotations, reports, notes, folders, generated artifacts, direction feedback, and other text or settings entered into the application.
- AI configuration and credentials: selected providers, models, evaluators, graders, orchestration settings, and account-provided AI-provider API keys. Stored provider keys are encrypted; the Service displays only limited status and key-hint information after storage.
- Usage and performance information: feature and run activity; model and provider identifiers; input and output token counts; estimated provider cost; latency; result and job status; monthly usage; error type; capacity measurements; and settings or feature preferences.
- Billing and transaction information: Stripe customer and subscription identifiers; plan; price and billing interval metadata; payment and subscription status; current period end; cancellation status; refund status and reason; billing-country results; and transaction timestamps. Stripe collects payment-card and billing details in its checkout and portal. Aplum AI does not receive or store full payment-card numbers.
- Security and compliance information: session and CSRF data; hashed, expiring verification and password-reset tokens; authentication and rate-limit events; IP address and request information available to the application or infrastructure; access, error, and security logs; and content-free records that identify the category, surface, action, and time of a restricted-content rejection. We do not intentionally retain the rejected text in those restricted-content event records.
- Communications: availability, verification, reset, security, support, privacy, and account communications; delivery identifiers and status; bounce, delay, failure, or suppression metadata; and records needed to prevent duplicate fallback messages.
- Local browser information: model preferences, pending-job references, collapsed or visible interface state, feature preferences, and similar application state stored locally in the browser.
The current customer Service does not accept ordinary server-side file uploads or per-prompt attachments. For the battery CSV import, the browser reads the selected CSV locally, displays normalized permitted rows for review, and transmits only the structured row values the user confirms. Aplum AI does not receive or store the original CSV bytes, local path, or original filename through that flow. Pasted or manually entered text is Customer Content even if it originated in another document.
3. Sources of information
We receive information directly from users, interested visitors, Business Customers, and Individual Customers; automatically from the browser, application, and hosting infrastructure when the Service is used; from Stripe about checkout, subscriptions, payments, cancellation, and refunds; from transactional-email providers about message delivery; from selected AI providers when they return outputs or usage information; and from a person who contacts us or validly directs us to act.
The current customer Service has no data-broker or purchased-consumer-profile intake.
4. Notice at collection and how we use information
At or before account registration, Aplum AI links to this Policy and identifies the account and eligibility information being requested. We collect those categories to create and secure the account, record the selected customer type and legal acceptance, provide the no-card trial when offered, communicate about the account, and administer access. We retain them using the criteria in Section 10. We do not sell them for money or share them for cross-context behavioral advertising.
We use information to:
- create, verify, authenticate, secure, and administer accounts;
- record and respond to a visitor's request to hear about Aplum availability;
- determine eligibility, plan access, feature limits, and capacity availability;
- provide the model tests, chats, evaluations, reports, exports, personalization, and other workflows a user requests;
- route the necessary content and instructions to selected target, evaluator, grader, image, or agent models;
- store and display histories, results, reports, settings, and enabled personalization or memory features;
- process subscriptions, reconcile signed Stripe events, provide billing controls, prevent duplicate refunds, and keep financial records;
- deliver verification, reset, service, security, and support communications and operate delivery fallback;
- screen for prohibited sensitive content, enforce restrictions, prevent abuse and fraud, investigate security events, and protect accounts, systems, users, and others;
- monitor reliability, troubleshoot errors, measure plan usage, manage capacity, and improve operational performance without using Customer Content for advertising;
- respond to support, privacy, legal, and intellectual-property requests;
- comply with law, establish or defend legal claims, and enforce our Terms; and
- carry out a financing, reorganization, acquisition, or sale subject to appropriate protections.
We do not use the substance of Customer Content to advertise to a user. Research use is governed separately as described below.
5. AI providers and provider keys
Aplum AI supports workflows involving Anthropic, OpenAI, Google, Sakana AI, xAI, and supported Meta-hosted model endpoints. The available list may change. A workflow may send relevant Customer Content and instructions to more than one provider—for example, a target model and a separate evaluator, grader, report, image, search, innovation, or analyst model. The interface and model configuration determine which providers are used.
Current production customer accounts use provider keys supplied for that account; customer activity is not charged to Aplum AI's internal provider keys. We encrypt stored customer provider keys using dedicated application key material and transmit a key to its provider as needed to authenticate a requested call. A user may replace or delete a stored key through account controls. The provider receives the request under the provider account associated with that key, and the customer's agreement with that provider may govern charges, retention, training, abuse monitoring, and other processing. Aplum AI does not control those provider practices.
We may introduce an expressly identified platform-funded provider feature in the future. We will update the Service or this Policy as appropriate if that materially changes the processing described here.
6. Research Mode
Research Mode is an Aplum AI administrative configuration; it is not currently a customer-accessible research console. When Research Mode is off, no new content-based Research Use begins, research-discounted enrollment is unavailable, and Aplum AI does not use the substance of Customer Content for generalized benchmark development, model training, advertising, or content-based product research. Processing needed to perform requested features, secure the Service, provide support, and measure non-content operational performance continues.
If Research Mode is enabled, Customer Content may be used for the specific research purposes and categories described in the then-current Research Mode Terms and Research Use Disclosure only after an eligible Business Customer's authorized account owner separately accepts those documents and affirmatively authorizes Research Use. Individual Customer Content is not eligible for Research Use under the current legal package. General Terms or this Policy alone do not grant that authorization. Withdrawal stops new content-based Research Use as described in those documents. Provider credentials, authentication information, billing data, and support communications are not authorized for Research Use.
We may preserve lawfully created aggregated or de-identified findings that are not reasonably linkable to a customer and do not reveal identifiable Customer Content. We do not attempt to re-identify de-identified material except to test safeguards or as required by law.
7. Disclosures and service providers
We disclose information only as reasonably necessary for the purposes above, including to:
- Fly.io and its infrastructure providers for application hosting, networking, storage, managed PostgreSQL, logs, and operational services;
- Stripe for checkout, payment processing, subscription management, invoices, customer billing controls, fraud prevention, cancellation, and refunds;
- Resend for primary transactional email and Google Workspace/Gmail SMTP as a delivery fallback;
- the AI providers selected or required by a configured workflow, currently including Anthropic, OpenAI, Google, Sakana AI, xAI, and supported Meta-hosted endpoints;
- professional advisers, contractors, and vendors who need information to provide legal, accounting, security, support, or operational services and are subject to appropriate duties;
- a government authority, court, regulator, or other party when disclosure is required by law or reasonably necessary to protect rights, safety, accounts, or systems; and
- a potential or completed investor, lender, acquirer, successor, or transaction adviser in connection with a financing, reorganization, merger, acquisition, or sale, subject to appropriate confidentiality and privacy protections.
The public site loads fonts from Google Fonts. As a result, a visitor's browser may make a direct request to Google that includes ordinary network information such as IP address and browser headers.
Aplum AI personnel with administrator access may access or, when reasonably necessary, impersonate a customer account to provide requested support, investigate security or billing issues, administer access, enforce restrictions, or comply with law. Authorized research personnel may access authorized research material only when Research Mode and valid Research Use Authorization permit it. We limit such access by role and purpose; administrator access is not a customer collaboration feature.
8. Cookies and local browser storage
We use a secure session cookie to maintain authentication and a CSRF cookie to protect authenticated requests. Production session cookies are configured as Secure, HttpOnly, and SameSite=Lax; the CSRF cookie must be readable by the application script and is configured as Secure and SameSite=Lax. We use local browser storage for interface preferences, selected model settings, pending-job references, and related requested functionality.
We do not currently use third-party advertising pixels or a third-party customer-behavior analytics script. Infrastructure and linked providers may process ordinary network information under their own policies. Browser settings can clear cookies and local storage, but doing so may sign the user out, remove preferences, or interrupt pending-job display.
9. Sale, targeted advertising, and browser privacy signals
We do not sell personal information for money. We do not share personal information for cross-context behavioral advertising, use Customer Content for targeted advertising, or knowingly sell or share personal information of anyone under 18. We do not offer financial incentives for personal information.
Because we do not engage in those sale or targeted-advertising practices, a Global Privacy Control signal does not change such processing. We will treat a recognized opt-out signal as required if our practices change or applicable law requires a response. We do not track users across unrelated websites for advertising, so browser "Do Not Track" signals do not cause a separate change in the Service.
10. Retention, deletion, and account closure
Retention depends on the type of information, account status, requested features, security and fraud needs, provider behavior, financial and legal obligations, and whether information is needed to resolve a dispute or enforce an agreement. We apply the following criteria:
- Account information, Customer Content, encrypted provider keys, histories, and generated artifacts generally remain in the active Service while the account is open or until the user deletes an item, an administrator clears the account data, or an approved deletion request is completed.
- Availability-interest information remains while reasonably needed to respond about Aplum access, manage duplicate requests, document the request, or until an approved deletion request is completed, subject to legal, security, or dispute-related retention needs.
- Verification and reset tokens expire and are single-use. Security, rate-limit, restricted-content, job, and delivery records remain only as long as reasonably needed for their operational, abuse-prevention, or evidentiary purpose.
- Stripe and Aplum AI retain transaction, subscription, cancellation, refund, tax, and fraud records under their respective legal, accounting, dispute, and operational requirements.
- If an account is permanently deleted after linked subscriptions are fully terminated, the active Aplum AI database removes the user account, Customer Content, stored provider keys, account tokens, usage records, and associated application data. Aplum AI retains a content-free reconciliation record containing internal deleted-account and deleting-administrator identifiers, Stripe customer and subscription identifiers, plan and terminal subscription status, subscription-period information, and account/deletion timestamps. This record supports transaction reconciliation, fraud prevention, disputes, and legal or financial obligations; it does not contain the deleted email address, password, provider keys, or Customer Content.
- Information may remain for a limited time in infrastructure logs, provider systems, or backup and disaster-recovery copies until those systems rotate or are securely overwritten. Restored data remains subject to the deletion or withdrawal decision and must not be returned to ordinary active use.
- Aggregated or de-identified information may be retained when it is not reasonably linkable to a customer and does not reveal identifiable Customer Content.
Account deletion does not itself cancel an active Stripe subscription. A linked subscription must be completely canceled before permanent account deletion so that future charges and transaction ownership can be reconciled safely. Users can cancel through the Stripe billing portal available in the Service or contact us for help. Aplum AI may retain information longer when required by law, legal hold, security investigation, fraud prevention, or an unresolved transaction or claim.
11. Security
We use safeguards designed for the nature of the information in our control. Confirmed measures include TLS in production; password hashing; hashed, expiring, single-use account tokens; encrypted provider keys with dedicated key material; secure session settings; CSRF protections; same-origin restrictions; security response headers; signed Stripe and email webhooks; access controls; plan and ownership checks; and restricted-content screening. No online system is perfectly secure, and we cannot guarantee that information will never be accessed, lost, altered, or disclosed improperly.
Users are responsible for protecting credentials, choosing appropriate provider settings, reviewing who can access their account email and provider accounts, and avoiding restricted information. Contact founder@aplumlabs.com promptly if unauthorized access is suspected.
12. Choices and privacy requests
Users can update available account settings and provider keys, delete stored provider keys, delete certain application items, manage or cancel subscriptions through billing controls, and withdraw Research Use Authorization when that feature is available. A person who submitted availability-interest information may ask us to access, correct, or delete it by contacting us. Because available exports cover particular product artifacts and summaries rather than every account or interest record, a person may request a broader copy of personal information by contacting us.
Subject to applicable law and appropriate verification, a user may request access to, correction of, or a portable copy of personal information; deletion; information about categories and recipients; restriction or objection where applicable; limitation of sensitive-personal-information use where applicable; or an appeal of a denied privacy request. A user may also ask us to confirm that we do not sell or share personal information for targeted advertising. We will not unlawfully discriminate for exercising a privacy right.
Send requests to founder@aplumlabs.com with the subject "Privacy Request." Describe the account and request without including passwords, provider keys, restricted data, or unnecessary Customer Content. We may verify the requester's identity, email ownership, business authority where applicable, and authorization of an agent. If Aplum AI processes personal information only on behalf of the requester's business, we may direct the request to that business. We may deny or limit a request where permitted by law and will explain available appeal steps. To appeal, reply with the subject "Privacy Appeal."
13. Children and restricted information
The Service is for users at least 18 years old, whether acting for a United States business or using an enabled Individual Customer account. We do not knowingly collect personal information from children through the customer Service. Contact us if a child is believed to have provided personal information.
Do not submit medical records; health, medical, wellness, diagnosis, treatment, or health-inference information about an identifiable person; information about minors; payment-card data; government identifiers; authentication secrets other than supported provider credentials; biometric identifiers; student records; nonpublic consumer-credit information; or information for regulated credit, housing, insurance, employment, or similar high-impact decisions. Technical acceptance is not permission. The Service may reject restricted submissions without retaining the submitted text.
14. United States operation
Aplum AI is operated from the United States and the public Service is limited to United States Business Customers and, when separately enabled, adult United States Individual Customers. Information is processed in the United States and may be processed in other locations where a selected provider or service provider operates. Do not use the Service from another jurisdiction if that processing or the applicable United States customer terms are not permitted.
15. Changes to this Policy
We may update this Policy prospectively. We will post the new version and effective date and provide additional notice when appropriate to the significance of the change. We will not use a Policy update alone to authorize materially broader content-based Research Use; that requires the separate acceptance and affirmative authorization described above.
16. Contact
Aplum AI LLC
Privacy email: founder@aplumlabs.com
For the safest response, do not put passwords, provider keys, payment-card data, medical information, or other restricted data in an email request.