PICKLE SENSEI — PRIVACY POLICY Last updated: September 11, 2026 This Privacy Policy explains how Raunak Gengiti, an individual who provides and operates the service under the name Pickle Sensei ("Pickle Sensei," "we," "us," or "our"), collects, uses, discloses, retains, and protects information when you use the Pickle Sensei mobile application, its account services, and the public pages served with those services (collectively, the "Service"). The most important privacy fact is simple: raw court video, camera frames, audio recorded with a clip, and body-pose landmarks stay on your device. Stroke analysis runs on your device. Pickle Sensei does not upload those raw media or pose-landmark files to its servers. 1. SCOPE AND RESPONSIBILITY This policy applies to information processed by Pickle Sensei through the Service. It does not replace the privacy notices of Apple, Google, RevenueCat, Supabase, YouTube, Vimeo, or another third party whose service you choose to use. Those parties process some information under their own terms and privacy policies, as described below. For purposes of applicable privacy law, Raunak Gengiti is the operator and, where that terminology applies, the business or data controller responsible for Pickle Sensei. The mailing address is 6737 Elegante Way, San Diego, California 92130, United States. 2. INFORMATION WE COLLECT OR PROCESS A. Account and sign-in information When you sign in with Apple or Google, we receive the information that the provider makes available and that is needed to create or access your account. This may include your email address, display name, sign-in provider, provider account identifier, and an internal Pickle Sensei account identifier. Apple or Google may not supply a display name. Any provider display name is separate from the name you enter during onboarding. The Google Sign-In software included in the app declares that it may process name, email address, phone number, network-derived coarse location, user and device identifiers, other authentication data, and usage data for sign-in, security, service operation, and analytics. Pickle Sensei requests only the basic sign-in identity needed to authenticate your account and does not request device location permission for Google sign-in. Google controls its own provider-side processing under its privacy notice. The provider identity token is used to establish a session and is not stored on your device by Pickle Sensei. A refresh credential is stored in the device's protected Keychain or Keystore so you can remain signed in. Access credentials are held temporarily while the app runs. Authentication records and server sessions are handled through Supabase Auth. B. Coaching profile and preferences Onboarding requires a name for account personalization. You may enter your preferred name or a nickname; we do not require or verify your legal name for this field. We use it to address you in the app. We also collect onboarding answers about skill level, dominant hand, training goals, and your biggest problem, and use them to set your starting focus area and personalize coaching, recommended drills, and the in-app experience. The questionnaire is required. You may choose "Prefer not to say" for gender. Completed answers are stored on your device before sign-in and linked to your account and synced to our servers after you sign in. C. Stroke analysis, activity, and progress information The app creates structured records from on-device analysis. Depending on the feature used, these records may include: • declared and detected stroke type; • camera view and capture mode; • capture and session timestamps, clip duration, frame rate, and dimensions; • technique scores, confidence values, movement measurements, phase timing, checkpoint results, limiting factors, and coaching recommendations; • app, model, detector, and scoring-version identifiers; • practice-session counts, completed drills, saved drills, streaks, achievements, rank, and progress summaries; and • whether a result was scored or the system abstained because confidence or evidence was insufficient. These structured records are stored locally. When you use a signed-in, synced account, some records needed for account history, progress, rank, allowance enforcement, and service continuity are also sent to our servers. They do not contain the raw video, raw audio, camera frames, or pose-landmark file from which the analysis was produced. D. Feedback and communications If you rate an analysis as accurate or not quite right, we collect that choice and, when applicable, the issue category. If you contact us, we receive the information in your message and the contact information you use. Before deleting an account, you may complete an optional exit survey. It can include a selected reason, what might have changed your decision, optional free-text comments, platform and app version, sign-in provider, approximate account age, whether the account had Pro access, and the number of scored analyses. Skipping the survey never prevents deletion. E. Consent and privacy-choice records We keep an auditable record of privacy choices made in Data & consent, including the scope, grant or withdrawal, policy or consent version, timestamp, source screen, platform/OS label, and capture-mode label. This lets us honor the current choice and demonstrate that optional processing was not enabled without an affirmative action. F. Optional model-improvement and evaluation records Optional model-improvement permission is off by default. If you turn it on, analysis feedback you submit and the associated structured analysis record may be marked eligible for review and model improvement. Turning it on does not upload raw video, raw audio, camera frames, or pose-landmark files. Evaluation telemetry is a separate, opt-in category and is also off by default. If an evaluation-telemetry control is offered and you enable it, we may collect a record of an analysis attempt: generated trial/capture and analysis identifiers, timestamps, declared stroke, result or abstention, latency, app/model versions, confidence and scoring claims, phase/contact timing, limiting factors, user flags, capture dimensions, and the consent version. These records are linked to your Pickle Sensei account while the account exists; they are not anonymous. They do not include the raw clip or pose-landmark file. Withdrawing an optional permission stops new records from becoming eligible or being collected under that permission. It does not retroactively undo processing already completed before withdrawal. Consent-history records are retained with the account for accountability and are deleted with it. G. Purchase and entitlement information Apple's App Store processes purchases. RevenueCat helps us validate receipts and determine access. We and RevenueCat may process your Pickle Sensei account identifier, store, product identifier, purchase and renewal history, transaction/receipt status, entitlement status, expiration date, trial or introductory-offer status, and related store metadata. We use this information to provide Pro features, prevent purchase fraud, restore purchases, support customers, and understand subscription performance. We do not receive or store your full payment-card number, bank-account details, or store-account password. Apple handles the payment. H. Network, app, and security information For offline ratings, the app creates a random installation identifier in the device's protected Keychain or Keystore and registers it with our server. We associate that identifier and a server-issued device identifier with your account, offline rating allocations, grant expiration times, and settlement receipts. We use these records to enable offline ratings, reconcile results when you reconnect, and prevent an allocation from being spent twice. These identifiers are not advertising identifiers and are not used for cross-app tracking. Retention after account deletion is described in Sections 7 and 8. When the app communicates with our API, servers necessarily process request information such as IP address, request time and route, app version, platform, authentication status, and response/error information. We use this data to deliver the Service, protect accounts, investigate errors, prevent abuse, and apply rate limits. Short-lived authentication caches, token hashes, and rate-limit counters normally expire within ten minutes. Hosting providers may retain infrastructure and security logs under their own documented schedules. We do not use a third-party advertising SDK, do not request the advertising identifier, and do not track activity across other companies' apps or websites for advertising. When you choose an externally hosted drill video, the embedded YouTube or Vimeo player may process the selected video or page, playback interactions, device or browser identifiers, network-derived coarse location, and information about advertisements displayed by the provider. The provider may show its own advertisements. YouTube embeds use its privacy-enhanced mode where supported, which Google states prevents the embedded view from being used to personalize the viewer's YouTube browsing experience or advertising outside the app. Pickle Sensei does not receive a provider advertising profile and does not use this information for cross-app tracking. I. Information kept only on your device The following is processed or stored locally by the app and is not uploaded to Pickle Sensei's servers: • raw recorded or imported video, camera frames, clip audio, and pose landmarks; • the contents of your photo library other than the single video you choose through the system picker; • notification preferences and locally scheduled reminder state; • device-local playback and interface state; and • protected sign-in credentials described in Section 2.A. Camera, microphone, photo-library selection, and notification access are used only after the relevant feature or permission is requested. You can change system permissions in device Settings, although disabling a required permission may prevent that feature from working. 3. INFORMATION WE DO NOT COLLECT FOR OUR SERVERS Pickle Sensei does not collect for its servers: • raw court video, raw camera frames, raw clip audio, or pose-landmark files; • payment-card or bank-account numbers; • precise or coarse GPS location; • address-book contacts; • advertising identifiers or cross-app advertising profiles; • face templates, face recognition, or identity inferred from body movement; or • medical records. We do not sell personal information. We do not share personal information for cross-context behavioral advertising, and we do not serve targeted ads. 4. HOW WE USE INFORMATION We use information described above to: • create, authenticate, secure, and support accounts; • provide on-device stroke analysis and sync structured results; • personalize coaching, drills, progress, streaks, achievements, and rank; • enforce free-analysis allowances and provide paid features; • process, verify, restore, and troubleshoot purchases; • remember and enforce consent and privacy choices; • respond to questions, feedback, and account requests; • operate, secure, debug, and improve the Service; • conduct optional evaluation or model-improvement work only within the permission you granted; and • comply with law, resolve disputes, and enforce our Terms of Use. Pickle Sensei uses automated, on-device processing to generate technique scores and coaching suggestions. These outputs do not make legal, employment, credit, insurance, medical, or similarly significant decisions about you. 5. WHEN WE DISCLOSE INFORMATION We disclose information only as needed for the purposes in this policy: • Supabase provides authentication, database, and Edge Function hosting for account and synced-service data. • RevenueCat receives an internal account identifier and purchase-related information to validate purchases, manage entitlements, prevent fraud, provide subscription analytics, and support purchase restoration. • Apple and Google provide sign-in and store-payment services. Information sent to or received from them is also governed by their policies. • Upstash may provide short-lived cache and rate-limit infrastructure when enabled. Cached values are derived service state, token hashes, and short-lived request counters rather than raw court media. • YouTube or Vimeo may receive ordinary web-request information, such as IP address, device/browser identifiers, network-derived coarse location, cookies or similar storage, the video or page viewed, playback interactions, and advertising data when you choose to open or play an externally hosted drill video. YouTube embeds use its privacy-enhanced domain where supported, but the video provider remains an independent third party and may show its own advertisements. • Professional advisers, authorities, or other parties may receive information when reasonably necessary to comply with law, protect rights and safety, investigate fraud or security incidents, or establish or defend legal claims. • If the Service is involved in a merger, financing, acquisition, reorganization, or sale of assets, information may be transferred subject to appropriate confidentiality and continued notice of applicable privacy practices. Service providers are permitted to process information only for contracted purposes and must protect it consistently with their agreements and applicable law. We do not permit them to use Pickle Sensei data for targeted advertising on our behalf. 6. LEGAL BASES FOR PROCESSING Where a law requires a legal basis, we rely on: • performance of a contract — to provide accounts, analysis history, personalization, purchases, and requested Service features; • consent — for optional model-improvement or evaluation processing and for device permissions where consent is the applicable basis; • legitimate interests — to secure, maintain, troubleshoot, and improve the Service, prevent fraud, and understand aggregate product performance, where those interests are not overridden by your rights; and • legal obligations and legal claims — when processing is needed to comply with law or establish, exercise, or defend rights. You may withdraw consent for future optional processing at any time without affecting the lawfulness of processing completed before withdrawal. 7. RETENTION We use the following retention rules: • Account, profile, synced analysis, training, progress, consent, feedback, and entitlement records are retained while the account is active and are deleted from our primary application database when account deletion completes, except as specifically stated below. • A pending deletion challenge expires after 15 minutes and stale challenge records are scheduled for deletion after expiration. • When you request account deletion, we keep a deletion operation record for 7 days after the request so that a deletion interrupted by a network or service failure can be resumed rather than repeated, so the app can show you the outcome for 24 hours, and so we can demonstrate that the deletion was carried out. This record holds the internal account identifier, the timestamps of each step, the outcome of the Sign in with Apple and RevenueCat clean-up steps, and any error codes recorded along the way. It does not contain your email address, name, clips, or analysis data, cannot be used to restore the account, and is restricted to service administration and security purposes. It is scheduled for deletion when the 7 days end. • Short-lived cache and rate-limit records normally expire within ten minutes. • RevenueCat webhook audit records are scheduled for deletion after 90 days. They are restricted to service administration and security purposes. • When RevenueCat reports that a store purchase was transferred from one Pickle Sensei account to another, we keep a purchase transfer reconciliation record: the transfer identifier, the internal account identifier of each account involved, RevenueCat's verified entitlement answer for each of them, and an append-only audit trail of each reconciliation step with timestamps. Because it documents which account a purchase was moved from and to, the record is not removed when one of those accounts is deleted; the deleted account is marked as no longer present instead. It contains no email address, name, clips, or analysis data, is not automatically deleted, and is kept for billing dispute resolution, fraud prevention, and security purposes only, restricted to service administration. • If you submit the optional exit survey and then delete your account, the account identifier is removed from that response during deletion. The resulting de-identified survey entry, including the coarse context listed in Section 2.D, may be retained for longitudinal product research until it is no longer useful. It is not used to contact you or recreate your identity. • To keep the free rating a once-per-person offer, we retain a one-way hash (SHA-256) of your sign-in provider's account identifier together with the number of scored analyses recorded under it. This free-rating record contains no email address, name, or Pickle Sensei account identifier and cannot be turned back into the identifier. It is used only to recognize that a sign-in identity has already used its free rating, is retained on the legitimate-interest basis of preventing free-tier abuse, survives account deletion for that reason, and is not used to contact you or to recreate your account. • When a free rating is allocated to your device for use while it is disconnected, we keep an append-only offline allocation record so that the allocation cannot be counted twice, cannot be reclaimed from a disconnected device, and can be recovered by the same installation. Each entry holds the internal account identifier, device and grant identifiers, the allocation ticket identifier, an installation key identifier, the same one-way (SHA-256) sign-in identity hashes as the free-rating record, whether the ticket was allocated, consumed for one scored analysis, or released, and a timestamp. It contains no email address, name, clips, or analysis data, cannot be turned back into your sign-in identifier, and survives account deletion so that the allocation follows the sign-in identity exactly as the free-rating record does. It is retained on the same legitimate-interest basis of preventing free-tier abuse and is not used to contact you. • Raw clips and other device-only data remain in app-private storage until removed by normal app cleanup or until you delete the app. Deleting your server account does not itself erase a clip file already stored on the phone. • Apple, Google, RevenueCat, video providers, and infrastructure providers may retain records under their own policies, including transaction, tax/accounting, fraud-prevention, and security records that they are legally required or permitted to keep. We may retain limited information longer when required by law, needed to complete a pending dispute or security investigation, or necessary to enforce legal rights. When possible, we restrict use during that period to the retention purpose. 8. ACCOUNT DELETION AND OTHER CHOICES You can delete a synced Pickle Sensei account in Settings → Manage account → Delete account. The optional survey can be skipped. After the separate final confirmation succeeds, deletion is permanent and removes the account and its associated profile, synced analysis history, sessions, progress data, saved drills, consent records, evaluation records, feedback, permits, rank state, and Pickle Sensei entitlement row from our primary application database. The backend also permanently deletes the customer record identified by the internal account identifier from RevenueCat. For an account created with Sign in with Apple, the backend revokes the stored Apple authorization before it deletes the account. An older account created before we began securely storing an Apple revocation credential may require you to remove Pickle Sensei manually in your Apple Account's Sign in with Apple settings; the app tells you when this exceptional step is needed. We fulfill the account deletion even if a legacy Apple credential is unavailable. A free rating you have already used is not restored by deleting the account. If you later sign in again with the same Apple or Google account, the hashed free-rating record described in Section 7 applies to the new account, so the free rating is not offered a second time. Deleting a Pickle Sensei account does not cancel an auto-renewing subscription managed by Apple, does not itself create a refund, and does not erase device-only clips. Cancel a subscription in the applicable store before deleting if you do not want it to renew. Apple subscriptions can be managed at https://apps.apple.com/account/subscriptions. You can also: • withdraw optional model-improvement permission in Settings → Data & consent; • change camera, microphone, photo, and notification permissions in device Settings; • manage or cancel subscriptions through your App Store account; and • ask us for access, correction, deletion, or a portable copy of personal information by emailing picklesenseidev@gmail.com. We may need to verify that a request relates to your account. Depending on where you live, you may also have rights to object to or restrict processing, withdraw consent, appeal a decision, or complain to a privacy regulator. We will not discriminate against you for exercising an applicable privacy right. 9. INTERNATIONAL PROCESSING Pickle Sensei and its providers may process information in countries other than the one where you live. Those countries may have different data- protection laws. Where required, we use contractual or other approved safeguards for cross-border transfers and honor mandatory local rights. 10. SECURITY We use administrative, technical, and organizational safeguards appropriate to the nature of the information. These include TLS encryption in transit, protected device credential storage, server-side authentication, scoped service credentials, per-user row-level database security, request-rate limits, input limits, and restricted server-only writes for entitlement and audit records. No storage or transmission system is completely secure, so we cannot guarantee absolute security. 11. CHILDREN The Service is not directed to children under 13, and we do not knowingly collect personal information from a child under 13. If you believe a child under 13 provided personal information, contact picklesenseidev@gmail.com so we can investigate and delete it. A parent or guardian is responsible for deciding whether an older minor may use the Service where local law requires consent. 12. CHANGES TO THIS POLICY We may update this policy as the Service, providers, or law changes. We will post the revised policy, change the date above, and provide additional notice when required. Material changes apply prospectively unless law permits or requires otherwise. 13. CONTACT Operator and privacy contact: Raunak Gengiti Mailing address: 6737 Elegante Way, San Diego, California 92130, United States Email for privacy questions and requests: picklesenseidev@gmail.com