하입덕 개인정보처리방침

HypeDuck helps creators, marketers, and e-commerce sellers discover and analyze public short-form content, including Instagram Reels, through the HypeDuck web dashboard and Chrome extension.

This Privacy Policy explains what information we collect, how we use it, and the choices available to users of HypeDuck.

Account information: We may collect information such as your name, email address, and account identifiers when you create or use a HypeDuck account.

Website content: When you intentionally use the Chrome extension to scan supported pages, HypeDuck may collect public content signals such as Reel URLs, captions, thumbnails, public engagement metadata, and related page information needed for analysis.

Public post metrics: Separately from scans, while you browse Threads and Instagram with the extension installed and signed in, it sends HypeDuck the numbers it has already received for public posts (views, likes, comments or replies, reposts, quotes and shares) together with the post code, the author’s username, the posting time, the author’s follower count and, for Instagram, the post type (reel, video, photo or carousel). These records do not show who viewed a post: we do not store the viewer’s account, device or the page address being viewed with them. We do not store post text, photos or videos, and we only collect posts whose author account is confirmed public. The extension sends no additional requests to Meta for this and uses only values it has already received.

공개 글 숫자 지표: 하입덕 확장은 사용자가 스레드와 인스타그램에서 보는 공개 글의 숫자 지표(조회수, 좋아요, 댓글과 답글, 리포스트, 인용, 공유)와 글 코드, 작성자 아이디, 게시 시각, 작성자 팔로워 수, 인스타그램 글 형식(릴스, 동영상, 사진, 캐러셀)을 하입덕 서버에 모읍니다. 누가 봤는지 알 수 없도록 열람한 사람의 계정, 기기, 보던 주소는 함께 저장하지 않습니다. 본문, 사진, 영상은 저장하지 않고, 작성자 계정이 공개로 확인된 글만 모읍니다. 이를 위해 Meta에 요청을 더 보내지 않고 확장이 이미 받은 값만 씁니다.

User activity: We may process actions taken inside HypeDuck, such as scan requests, selected Reels, analysis jobs, dashboard interactions, and extension workflow state.

Technical information: We may collect logs, device/browser information, and diagnostic data needed to secure, debug, and improve the service.

We use information to provide HypeDuck’s core functionality: scanning supported Instagram Reels pages at the user’s request, sending selected content signals to the HypeDuck dashboard, and generating analysis results for content research and marketing workflows.

In addition to the scans you start, we use the public post metrics described above for HypeDuck features such as finding posts that are taking off and ranking posts.

확장이 모은 공개 글 숫자 지표는 사용자가 직접 실행한 스캔과 별도로 터진 글 찾기, 순위 같은 하입덕 기능에 씁니다.

We also use information to authenticate users, maintain extension and dashboard state, improve product reliability, prevent abuse, and provide support.

The HypeDuck Chrome extension requests permissions only to support its stated purpose. These permissions may include storage, activeTab, tabs, cookies, scripting, offscreen, alarms, debugger, unlimitedStorage, and host permissions for supported websites and HypeDuck services.

The extension does not use these permissions to monitor unrelated browsing activity, sell user data, or profile users for unrelated advertising purposes.

We do not sell user data. We do not use or transfer user data for purposes unrelated to HypeDuck’s single purpose, and we do not use user data for creditworthiness, lending, or similar eligibility decisions.

We may use service providers for hosting, databases, analytics, storage, authentication, and AI-assisted analysis. These providers process information only as needed to operate HypeDuck.

We retain information for as long as needed to provide the service, comply with legal obligations, resolve disputes, prevent abuse, and maintain business records.

Public post metrics are kept as the latest values for each post and deleted 180 days after the post was last observed; hourly records are deleted after 30 days. Cleanup runs in bounded batches and does not guarantee deletion immediately after a deadline.

공개 글 숫자 지표는 글마다 최신값을 두고 마지막으로 관측된 뒤 180일이 지나면 지우며, 시간대별 기록은 30일 뒤 지웁니다. 정리는 처리량이 제한된 배치로 진행되어 기한 직후 삭제 완료를 보장하지 않습니다.

We use reasonable administrative, technical, and organizational safeguards to protect information, but no online service can guarantee absolute security.

You can stop using the Chrome extension, remove it from your browser, or contact us to request assistance with account or data-related questions.

Where required by applicable law, you may request access, correction, deletion, or export of certain personal information.

If you choose to connect your own Instagram professional account or Threads account, HypeDuck uses secure OAuth login (Instagram API with Instagram Login, and the Threads API) and receives an access token for that account.

Where enabled by HypeDuck’s access policy and the permissions you grant, the connected Instagram account supports the features below. Availability may be restricted to designated review accounts; implementation does not mean that every permission has Meta approval for general users.

instagram_business_basic lets HypeDuck read your connected account profile (including account ID, username, name, profile picture, account type and public counts) and media information to identify the account and display your media. Profile information and recent post captions also help build a content persona that personalizes generated scripts and captions.

instagram_business_content_publish lets HypeDuck publish the media and caption you select and confirm to your connected Instagram account. instagram_business_manage_comments lets HypeDuck read comments on media owned by that account and request replies, hiding or deletion of comments that you confirm. The server checks the media and comment ownership; comment deletion does not mean deletion of Instagram posts. instagram_business_manage_insights lets HypeDuck retrieve and display account and owned-media statistics.

If you explicitly enable a keyword resource rule for a selected post or Reel, instagram_business_manage_comments also lets HypeDuck check new comments and send one private reply containing your configured text and resource link when the keyword matches. The server verifies the connected account, media ownership and comment age. Private replies are limited to one per comment within seven days; the rule does not initiate further messages. Rules are saved disabled until you enable them.

Keyword rules store the selected media ID, keyword, reply text, resource URL and enabled state. Delivery records store comment and media identifiers, status, error code and timestamps. We keep a short-lived hashed account-and-comment reservation to prevent duplicate private replies even after reconnecting; it does not contain the reply text. Disconnecting removes connection-linked rules and delivery records; the duplicate-prevention reservation expires separately.

instagram_business_manage_messages lets HypeDuck process conversation, participant and message identifiers, message text and timestamps to display DM conversations and send replies you explicitly write and confirm, and, only if you switch it on, the keyword auto-reply described below. Before a manual reply, the server checks the current connected account and permissions and verifies evidence of an incoming message from the recipient within the last 24 hours. An optional human-agent reply is off by default and opens only for review-allowlisted accounts or after Meta approves the Human Agent feature: a reply you personally write and confirm may then be sent with Instagram's HUMAN_AGENT tag after 24 hours and before 7 days have passed since the recipient's latest message, and the server re-checks that window and the same permissions immediately before sending. Automated, AI-generated or promotional messages are never sent this way. A reply request does not guarantee delivery or a read receipt.

The DM reply operation ledger stores request identity, the reply kind (standard or human-agent), a request hash, state and provider result identifiers without saving the reply body as its operation payload. Browser tab recovery stores only a request key. These statements describe those specific stores, not every place where message text may be processed. Webhook reply-window records hold connected-account, peer and message identifiers, recent incoming-message time and expiry; retention cleanup removes expired records in bounded batches, without promising immediate erasure when the reply window closes.

If you explicitly enable a keyword auto-reply for direct messages, instagram_business_manage_messages also lets HypeDuck compare the text of new incoming text messages with your keyword and send the single text reply you wrote. It applies only to messages received after you switched it on and within the last 24 hours, never uses AI-generated text, sends at most one automatic reply per minute and five per day to the same person, and does not guarantee delivery or a read receipt. A message containing the word 상담원 (“agent”), a reply you send yourself, or the person-handling button stops automatic replies for that conversation until you resume them. Replies you write directly in the Instagram app are not detected, so use the button in that case. The rule is saved disabled until you enable it.

The keyword auto-reply stores your rule (keyword and reply text), delivery records (status, the other person’s Instagram ID, error code and timestamps, without their message text), a one-way hash of the account and message ID for about 25 hours to prevent duplicate replies, and the human-handling state (connected-account ID, the other person’s ID and timestamps). Delivery records are removed after 30 days and idle automatic-mode state after 2 days. Human-handling state is kept while the account has a connection so that automation never overrides a person; it is removed 30 days after cleanup first finds the account without any connection, and immediately when a verified Meta deletion request or your account withdrawal leaves the account without any connection. Disconnecting removes your rule and delivery records.

If you schedule a post, instagram_business_content_publish lets HypeDuck publish the photo, Reel, carousel or Story and the caption you chose to your connected Instagram account at the time you set, from 2 minutes to 30 days ahead. Creating the schedule is your authorization: HypeDuck does not ask again when the time comes. Just before publishing it re-checks the connection, account type and permissions and does not publish if they no longer allow it, and it does not publish if more than 15 minutes have passed. You can cancel until the publish request has been sent and change the time before processing starts; to change the content, cancel and schedule again. If the outcome cannot be confirmed, HypeDuck does not send the same request again and shows that you need to check Instagram. Story schedules send no caption or AI label, and Stories are available only to accounts whose account-type eligibility is confirmed and to designated review accounts.

When you schedule a post, HypeDuck copies the file you selected into its cloud storage as an immutable copy and publishes from that copy rather than from the original; Instagram retrieves the file from that storage address when the post is published. The copy’s path contains your user ID, a request ID and a random attempt ID, and just before publishing HypeDuck checks that the copy still matches what it copied (size, type and version fingerprint). For each schedule we store the kind, the time and time zone you chose, the caption and AI-generated flag, your HypeDuck user and connection IDs, a snapshot of the Instagram account ID, app ID, app-scoped user ID, account type and granted permissions, request keys and one-way hashes, the copy’s location, size, type and fingerprint, status and error codes, processing, cancel and publish-request timestamps, and the Instagram result ID once published. Related Instagram operation and Story status records, and a separate cleanup reference (user and connection IDs, the Instagram app ID and app-scoped user ID at the time it was made, request key, cleanup state and timestamps, no file content), are stored too.

We delete the stored copy only after the outcome is confirmed: Instagram confirmed the post, Instagram reported that the media container failed or expired so nothing was published, or the schedule was canceled, missed or blocked and no creation request had been sent. A cleanup job deletes the copy at least 24 hours after confirmation, and deletes the schedule record 30 days after it ended once the copy is gone. If the outcome cannot be confirmed (a timeout, an unknown result, or a cancellation or block after a request may have been sent), the copy and the record are kept without an automatic expiry until the outcome is confirmed, an operator closes it, or a deletion request is processed; where HypeDuck knows the media container ID it re-reads its status once an hour, at most 24 times, to try to confirm it. Temporary copies that never became part of a schedule are removed no sooner than 2 hours after they were made, with repeated checks for late-arriving copies. Cleanup runs in bounded batches and does not guarantee deletion immediately after a deadline.

Disconnecting an account, deleting a linked account, withdrawing your account and a verified Meta deletion request delete your schedule records and the related Instagram operation records, and in the same step first mark that user’s (or that connection’s) stored copies for deletion; a verified Meta deletion request reaches the copies by the Instagram app ID and app-scoped user ID recorded when they were made, even if you had already disconnected. If the mark cannot be written, the connection or account is not deleted. Once marked, the cleanup job removes the copies even when the outcome was never confirmed, and no new copy is started. A copy that was already being made can finish late, so the job checks again every 2 hours, at most 48 times, until it is established that copying has ended; if that cannot be established it keeps the mark and reports it rather than treating elapsed time as completion. Cleanup runs in batches and does not guarantee deletion immediately, and a Meta deletion receipt stays at “additional review needed” until the stored copies are actually erased. Older references that did not record those IDs are never matched to a Meta request by guesswork and stay flagged for review. Requests already sent to Instagram and posts already published cannot be undone, and this is not a function for deleting Instagram posts.

If you separately connect an Instagram professional account through Facebook Login (a connection independent of Instagram Login that does not move your account), HypeDuck uses instagram_basic, pages_show_list and pages_read_engagement to read your Facebook user ID, the Pages you can access with your roles, the Instagram account linked to the Page you choose, and the list of your own posts and Stories. Where enabled by HypeDuck’s access policy and the permissions you grant, instagram_manage_contents lets HypeDuck delete one of your own non-ad posts, Stories or Reels after you pick it and confirm; deleting a carousel deletes the whole album, the deletion cannot be undone, and an unclear result is shown as needing review instead of being assumed. instagram_manage_engagement (with pages_read_user_content) lets HypeDuck send one like or unlike that you click on your own posts, on comments and replies on your own posts (listing comments uses instagram_manage_comments), and on posts of public business or creator accounts (finding those posts uses instagram_manage_insights). Likes are never bulk, scheduled or automatic, are not available for private accounts or Stories, and delivery is not guaranteed. Requests for the same Instagram account are sent one at a time, spaced apart, and an unconfirmed request is never silently repeated.

The Facebook Login connection stores your Facebook and Instagram account IDs, the Page name, the username, the permissions granted and token expiry times; the Facebook user access token is encrypted at rest. Temporary authorization records expire after 10 minutes and their encrypted token is erased when they expire. Delete and like request records keep a request key, the target ID, state, error code and timestamps, never post content or tokens. Disconnecting only the Facebook connection removes its token and pending authorization records, cancels requests that were never sent and leaves your Instagram Login connection untouched; records of requests that were already sent or whose result is unknown stay, without any token, so the same request is not repeated by mistake. A Meta deauthorization callback erases the tokens of every connection of that Facebook user. A verified Meta deletion request or your account withdrawal removes the connection, pending authorization records and request records. Finished authorization records and never-sent rejected request records are cleaned up after 30 days in bounded batches, which does not promise immediate erasure at the deadline.

For the connected Threads account, threads_basic lets HypeDuck display the account you connected. If you also grant threads_content_publish and threads_manage_replies, HypeDuck can publish a root post and the ordered tail replies you wrote and confirmed for that publication. This can run immediately when you confirm it or at the time you scheduled the already-approved content. Each tail is linked to the actual preceding post ID returned by Threads. HypeDuck does not use this permission to manage unrelated conversations or other users' content. See our Security & Data Behavior page for the full capability list.

If you grant threads_delete, you can remove your own Threads post that HypeDuck previously published through the official API. HypeDuck resolves the account and post ID from your HypeDuck publication record, does not accept an arbitrary Threads post ID, and requires a separate irreversible-action confirmation before requesting deletion. The HypeDuck record is then marked deleted so the action remains auditable.

Access tokens are encrypted at rest. We also store connection identifiers, profile details, granted permissions, token expiry and operation records. Connected-account data is never sold or shared for purposes unrelated to HypeDuck.

Disconnecting in Settings removes the stored connection and tokens; it does not automatically erase all authored content, personas or analysis results. A verified Meta deletion request returns a receipt and processing status for the scoped connection-data request; older or copied data may require further review. An email can request broader account-data deletion, but submitting the request is not confirmation that deletion has completed.

See our Data Deletion page for request options and status details.

If you choose to connect a publishing channel through HypeDuck’s own app (YouTube, TikTok, X, LinkedIn, Pinterest, Google Business Profile, a Facebook Page or Meta Ads), you sign in on that platform’s own consent screen and HypeDuck receives an access token, and a refresh token where the platform issues one, for the account you approve. HypeDuck never receives your password for that platform.

HypeDuck uses that access only to do what you ask in the HypeDuck dashboard or with your own HypeDuck API key: show which account is connected, publish or schedule the text, photos and videos you choose with the settings you choose (for example a YouTube title and privacy setting), check whether a post was published, and, where the platform allows it, delete a post you published through HypeDuck. For Meta Ads, HypeDuck reads your ad accounts, campaigns, ad sets, ads, lead form names and insights, creates new campaigns, ad sets or ads only when you request it and always in a paused state, changes a status only when you request it, and sends the product catalog items you choose to sync.

For YouTube, HypeDuck requests youtube.upload to upload the videos you choose and youtube.readonly to show the channel you connected. It stores the channel ID and title, the encrypted tokens and their expiry, and records of the uploads you requested. HypeDuck uses YouTube API Services: by connecting a YouTube channel you agree to the YouTube Terms of Service and acknowledge the Google Privacy Policy. You can revoke HypeDuck’s access at any time on your Google Account permissions page.

HypeDuck’s use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements. We do not sell Google user data, do not use it for advertising, do not use it to train AI models, and do not allow people to read it except with your permission, for security purposes, or to comply with law.

Access and refresh tokens are encrypted at rest and are never returned by the HypeDuck API. Media you attach is stored in HypeDuck’s cloud storage so that the platform can retrieve it when the post is published. Content you publish becomes subject to that platform’s own terms and privacy policy.

Disconnecting a channel in HypeDuck erases its stored tokens immediately, and scheduled posts for that channel are not published afterwards; posts that were already published stay on the platform. You can also revoke access in the platform’s own settings. To delete the remaining connection and post records, use the Data Deletion page or email support@hypeduck.ai.

If you have questions about this Privacy Policy or HypeDuck’s privacy practices, contact us at support@hypeduck.ai.