Privacy Policy
Announced: October 29, 2026 · Effective date: November 28, 2026
View the previous Privacy Policy (effective October 23, 2026)
View the table comparing this version with the previous one
1. General Provisions and Purposes of Processing
Uniquegood Company Inc. (the "Company") complies with the Personal Information Protection Act (Korea) and other applicable laws, and processes personal information for the following purposes: account authentication and member management; provision of the Service and execution of agent tasks; integration with external services the user connects (such as Google); fee payment and settlement (payments through Paddle); usage metering and quota management; security and prevention of fraudulent use; detection and analysis of service errors and recovery from outages and disasters; customer support; service improvement and statistical analysis; and confirming demand before the paid launch and, where the user has optionally left contact details, sending launch and adoption notices. The Company does not sell personal information or share it with third parties for advertising purposes.
2. Items of Personal Information Collected and Collection Methods
(Required for sign-up and sign-in) Email address and name (if no name is entered, the part of the email address before the @ is used instead). Sign-up and sign-in use an email verification code, Sign in with Google, or Sign in with Apple; with Sign in with Apple, the user identifier provided by Apple is also stored. No password is collected for new sign-ups; for accounts previously created with a password, the password is stored as an irreversible one-way hash. (Optional) Profile photo and display name; information received from a Google account the user connects (Section 15); API keys and subscription sign-in credentials for AI providers connected by the user, together with the email address of that account; and tokens for connected external services (credentials and tokens are stored encrypted with AES-256-GCM). (Required for the non-member pre-launch survey) Organization name, expected headcount, desired plan, and selected needs. (Optional for the non-member pre-launch survey) Detailed needs and an email address for launch notices. (Collected automatically in the course of using the Service) IP address, cookies, device, browser, and app information, access records, service usage records (channel visits, read status, search history, usage, etc.), audit logs, and (mobile apps) device tokens and installation identifiers for push notifications and Lock Screen live activities. (Error and crash reports) Error information when an error occurs on the server, the web, or the desktop app (error type, scrubbed message and stack, browser and operating system information, an opaque user identifier, etc.), and device information (device model, operating system, memory and storage size, jailbreak/root status, installation identifiers, etc.) and crash information when the mobile app terminates abnormally; the detailed items are listed in Sections 6-8 and 6-9. (Generated in the course of use) Content the user enters or uploads, such as workspaces, channels, messages, files, and task outputs; note recordings and their transcripts (including speaker separation and speaker names assigned by the user); ratings of and comments on AI answers; emails received at service addresses (sender, recipients, subject, body, and attachments); when a Slack export file is imported, the names and profile photos of the original authors; and customer inquiries. (Payments) Paid purchases on the AXIT website are processed by Paddle as the merchant of record, and the email address, payment method, and billing information that the user enters on the Paddle checkout are collected directly by Paddle. The Company does not store payment method details such as card numbers and keeps only Paddle customer and subscription identifiers, plan, billing cycle, number of seats, amount paid, and subscription status. Collection takes place through sign-up and sign-in, use of the Service, connection of external services, submission of the pre-launch survey, and automated collection tools (cookies and error and crash reporting tools).
3. Processing and Retention Period
In principle, personal information is retained until membership withdrawal and is destroyed without undue delay upon withdrawal. What is deleted and what remains when a user withdraws or deletes a workspace or content is set out in Section 7. Responses to the non-member pre-launch survey are kept for one year from submission and then destroyed automatically; if the respondent requests deletion or withdraws consent at contact@uniquegood.biz, the response is destroyed without undue delay after verification. However, the following records are retained for the applicable period as required by law and then destroyed: records on contracts and withdrawal of offers, 5 years; records on payments and the supply of goods, etc., 5 years; records on consumer complaints or dispute resolution, 3 years; records on display and advertising, 6 months (all under the Act on the Consumer Protection in Electronic Commerce, etc. (Korea)); and login and access records (access date and time, IP address, etc.), 3 months (Protection of Communications Secrets Act (Korea)). Backups and copies kept for service stability and disaster recovery are destroyed automatically after the following periods: automated database backups, 7 days (Seoul), and their copies in Tokyo, Japan, 7 days; file storage backups, 35 days (Seoul); and automated search server snapshots, 14 days (Seoul). The Tokyo replica of the file storage is kept in the same state as the original, so data deleted from the original is also deleted from the replica. Uploaded files that are deleted are also deleted from the upload storage (Amazon S3), and previous versions of files deleted or changed there are deleted automatically after 30 days in Seoul and after 14 days in the Tokyo copy. During these periods, information that has already been deleted may remain in backups and copies. Temporary copies of working folders moved to an isolated AI execution environment are deleted when that execution ends, and any copies that remain are deleted automatically after one day. Error report information (Sentry) and mobile app crash report information (Firebase Crashlytics) are deleted automatically after at most 90 days under each service's retention policy. Other logs are retained separately by type as follows: service usage logs (records of feature use) are retained until membership withdrawal, and audit logs for security and fraud prevention are retained for one year from collection and then destroyed. However, audit log entries that changed contract terms, such as plans and credits (grants, refunds, and revocations), are kept for 5 years as the contract records above. The original copies of emails received at service addresses are kept as data of the workspace to which they were delivered, and original copies not delivered to any workspace (other than mail sent to a service address currently in use) are destroyed one year after receipt.
4. Provision to Third Parties
The Company does not provide users' personal information to third parties without consent, except where there is a statutory basis or where an investigative authority makes a request in accordance with procedures prescribed by law. Where, at the user's instruction, task outputs or other information are sent to an external service that the user has connected (such as GitHub, Notion, Vercel, Cloudflare, or a bot event URL), the terms and privacy policy of that service apply.
5. Outsourcing of Processing
To the extent necessary to provide the Service, the Company engages the following service providers to process personal information: Amazon Web Services, Inc. (operation of the service infrastructure, such as servers, databases, file storage, search (including conversion into search vectors), logs, and backups — stored in the Seoul Region (Republic of Korea); storage of disaster-recovery copies — Tokyo Region, Japan; and AI model calls — Amazon Bedrock), Cloudflare, Inc. (DNS, content delivery, security, and network tunneling for some service addresses; United States), Apple Inc. (delivery of iOS push notifications and Lock Screen live activities; United States), Google LLC (web analytics — Google Analytics; Android push notifications — Firebase Cloud Messaging; mobile app crash reporting — Firebase Crashlytics; United States), Functional Software, Inc. (server and web error reporting — Sentry; United States), Paddle.com Market Limited (payment, subscription, and refund processing for the AXIT website as the merchant of record; United Kingdom), Mailgun Technologies, Inc. (sending and receiving email; United States), Eleven Labs Inc. (transcription and speaker separation of note recordings in workspaces designated by the Company; United States), and the AI providers in Sections 6-4 and 6-5 (AI execution, note transcription, library summaries, etc.). Details of processing outside Korea are set out in Section 6. Outsourcing agreements incorporate the protective measures required by applicable law, and this Policy will be updated without undue delay if any service provider changes.
6. Cross-Border Transfer of Personal Information
The Company stores service data in principle in the Republic of Korea (AWS Seoul Region), and personal information is transferred outside Korea as set out below in the course of providing the Service. In accordance with the Personal Information Protection Act (Korea), the Company discloses below the recipient and its contact details, the items transferred, the destination country, the timing and method of transfer, the purpose of use, the retention and use period, and the method, procedure, and effect of refusing the transfer. Of these, infrastructure and disaster recovery (Cloudflare and AWS), error and crash reporting (Sentry and Firebase Crashlytics), push notifications (Apple and Firebase Cloud Messaging), payments (Paddle), email (Mailgun), and the AI features that users instruct or use (Anthropic, OpenAI, Amazon Bedrock, Z.ai, xAI, Moonshot AI/Kimi, Google/Gemini, and ElevenLabs) constitute outsourcing and storage necessary for concluding and performing the contract. Web analytics (Google Analytics), by contrast, serves statistics and service improvement and is not essential to performing the contract; users may refuse it at any time by the methods in Section 6-2, and refusal does not affect use of the Service.
6-1 Cloudflare, Inc. (United States)
· Recipient and contact: Cloudflare, Inc. (privacy@cloudflare.com) · Items transferred: access IP address, device/browser information, and the data sent and received when some service addresses (such as www.axit.team) are accessed · Destination country: United States (and Cloudflare's global network) · Timing and method: routed and processed through the network (TLS-encrypted) whenever those addresses are accessed · Purpose: DNS, content delivery, security, and network tunneling · Retention and use period: processed for the duration of the outsourced service (provision of the Service); log data is kept briefly and then deleted under Cloudflare's policies · Method and effect of refusal: due to the nature of the infrastructure, individual refusal is not possible; users who wish to refuse may stop using the Service (in which case the Service cannot be used).
6-2 Google LLC (United States) — Google Analytics
· Recipient and contact: Google LLC (privacy inquiries: https://support.google.com/policies; privacy policy: https://policies.google.com/privacy) · Items transferred: cookie identifiers, visit and usage behavior, and device information · Destination country: United States · Timing and method: transmitted via cookies when the web analytics script runs · Purpose: web usage statistics and service improvement · Retention and use period: processed for the outsourcing period; Google Analytics user and event data is deleted automatically after at most 14 months · Method and effect of refusal: users may refuse by blocking cookies in the browser or by installing the Google Analytics opt-out browser add-on (https://tools.google.com/dlpage/gaoptout); refusal does not affect use of the Service.
6-3 Apple Inc. (United States)
· Recipient and contact: Apple Inc. (privacy@apple.com, https://www.apple.com/legal/privacy/contact/) · Items transferred: device tokens for iOS push notifications and Lock Screen live activities, and notification content (notification title, up to the first 180 characters of the message, the sender's name and profile photo URL, and, for live activities, the task title and progress steps) · Destination country: United States · Timing and method: transmitted through APNs when a notification or live activity is sent · Purpose: delivery of iOS push notifications and Lock Screen live activities · Retention and use period: while the device token is valid (until the app is deleted, the token is invalidated, the user signs out, or membership is withdrawn) · Method and effect of refusal: users may refuse by turning off notification permissions in device settings or in the app; refusal only stops notifications and does not otherwise affect use of the Service.
6-4 Anthropic PBC and OpenAI, LLC (United States)
· Recipient and contact: Anthropic PBC (privacy@anthropic.com) / OpenAI, LLC (privacy@openai.com) · Items transferred: prompts and content sent for AI execution and conversations (the user's instructions and the messages and file content needed for the task) and their outputs, and (OpenAI) note recordings for transcription and prompts for image generation · Destination country: United States · Timing and method: when a user instructs an execution or conversation, or requests a note transcription or image generation, and this provider handles that work under the workspace's AI settings, the data is transmitted at that time via API (TLS-encrypted), using a subscription account or API key connected by the user or the Company's key (paid with credits) · Purpose: AI execution in accordance with the user's instructions (performance of the contract) · Retention and use period: transmitted and processed for as long as necessary to handle the request; when processed via API, each provider keeps inputs and outputs for a period under its own policies (typically up to 30 days) for purposes such as abuse monitoring and then deletes them, and when processed through a subscription account connected by the user, the provider's terms, privacy policy, and training settings applicable to that account apply · Method and effect of refusal: users may disconnect the provider in settings or choose a different provider in the workspace's AI settings, and workspace owners and admins can choose "Allow Bedrock models only" so that AI executions and conversations run only on Amazon Bedrock (Section 6-7); refusal only limits AI features that use that provider and does not otherwise affect use of the Service.
6-5 Z.ai, xAI, Moonshot AI/Kimi, and Google (Gemini API)
· Recipient and contact: JINGSHENG HENGXING TECHNOLOGY PTE. LTD. (Z.ai, https://chat.z.ai/legal-agreement/privacy-policy) / X.AI LLC (xAI, https://x.ai/legal/privacy-policy) / Moonshot AI PTE. LTD. (Kimi, https://www.kimi.com/user/agreement/userPrivacy?version=v2) / Google LLC (Gemini API, https://policies.google.com/privacy) · Items transferred: prompts and content sent for AI execution and conversations and their outputs; (Z.ai) the file names, opening content, and outlines of files added to the library, for summaries, and the sentences entered when creating automations (Maky); (Google Gemini) library summaries of images and scanned documents, and the instructions and attachments for attachment descriptions and task-routing judgments piloted in workspaces designated by the Company · Destination country: Singapore (Z.ai and Kimi), United States (xAI and Google), and the processing locations disclosed in each provider's policies · Timing and method: transmitted at the time via API (TLS-encrypted) when a user instructs execution with that provider (using an API key or subscription account connected by the user, or the Company's key) and when the features above run, such as when a file is added to the library · Purpose: AI execution and provision of service features such as library summaries and attachment descriptions (performance of the contract) · Retention and use period: the period necessary to handle the request and to meet each provider's security and legal obligations, as set out in each provider's privacy policy and service agreement · Method and effect of refusal: users may disconnect the provider in settings or choose a different provider, and nothing is sent to these providers when workspace owners and admins choose "Allow Bedrock models only"; refusal only limits AI features that use that provider (such as library summaries).
6-6 Amazon Web Services, Inc. (Japan) — Disaster-Recovery Copies
· Recipient and contact: Amazon Web Services, Inc. (contact channel at https://aws.amazon.com/privacy/; 410 Terry Avenue North, Seattle, WA 98109-5210, USA) · Items transferred: copies of all personal information held in the service database and file storage (account information; user content such as workspaces, messages, files, and task outputs; encrypted connection credentials; service usage records; etc.) · Destination country: Japan (AWS Asia Pacific (Tokyo) Region, ap-northeast-1) · Timing and method: from October 16, 2026, replicated from the Seoul Region in encrypted form on an ongoing basis using AWS replication features (the database whenever an automated backup is created, and the file storage and uploaded files whenever they change) · Purpose: recovery of the Service and data in the event of a large-scale outage or disaster in the Seoul Region · Retention and use period: database automated backup copies, 7 days; the file storage replica is kept in the same state as the original (data deleted from the original is also deleted from the replica); previous versions of files deleted or changed in the upload copy are deleted after 14 days · Method and effect of refusal: because these copies hold the entire service data for recovery, individual refusal is not possible; users who wish to refuse may stop using the Service by withdrawing membership (in which case the Service cannot be used).
6-7 Amazon Web Services, Inc. — Amazon Bedrock (Global Cross-Region Inference)
· Recipient and contact: Amazon Web Services, Inc. (as in Section 6-6) · Items transferred: prompts and content sent for AI execution and conversations and their outputs (calls to Claude and GPT models processed with the Company's key and credits) · Destination country: requests are sent to the Seoul Region, but because global cross-Region inference is used, the processing itself may take place in any of the commercial AWS Regions worldwide (United States, Europe, Asia Pacific, etc.) · Timing and method: transmitted at the time of AI execution over the AWS network (TLS-encrypted) · Purpose: AI execution in accordance with the user's instructions (performance of the contract) · Retention and use period: processed only for as long as needed to handle the request; AWS states that Amazon Bedrock does not store inputs and outputs or use them to train models, and the Company does not turn on model invocation logging · Method and effect of refusal: when the user connects their own subscription account or API key in the workspace's AI settings, those executions do not go through Amazon Bedrock; refusal may limit AI execution with the Company's key (credits).
6-8 Functional Software, Inc. (United States) — Sentry Error Reporting
· Recipient and contact: Functional Software, Inc. d/b/a Sentry (compliance@sentry.io; 45 Fremont Street, 8th Floor, San Francisco, CA 94105, USA) · Items transferred: when an error occurs on the server or the web (including the desktop app), the error type, the error message with sensitive information removed, stack information (file, function, line), the service version (release) and environment, the pattern of the request path (not the actual address) and the request method, workspace and task identifiers, an opaque user identifier (u_…), browser and operating system information (User-Agent), and (web) the addresses (without query strings) and status of the screen navigations and network requests just before the error. Request bodies, cookies, email addresses, tokens, file names, and conversation content are removed before sending, and the service is configured not to store IP addresses · Destination country: United States · Timing and method: transmitted over TLS whenever an error occurs · Purpose: detection and analysis of service errors and ensuring service stability · Retention and use period: processed for the outsourcing period; error records are deleted automatically after at most 90 days under Sentry's retention policy · Method and effect of refusal: this processing is necessary for service stability and there is no option to refuse it individually; users who wish to refuse may stop using the Service (withdraw membership).
6-9 Google LLC (United States) — Firebase Crashlytics and Firebase Cloud Messaging
· Recipient and contact: Google LLC (https://firebase.google.com/support/privacy, https://support.google.com/policies) · Items transferred: (Firebase Cloud Messaging) device tokens for Android push notifications and notification content (title, up to the first 180 characters of the message, the sender's name and profile photo URL, and identifiers); (Firebase Crashlytics) when the mobile app (iOS and Android) terminates abnormally or reports an error, the device model, operating system name and version, CPU architecture, memory and storage size, jailbreak/root status, app version, installation identifiers (Crashlytics installation UUID and Firebase installation ID), stack information, scrubbed exception messages, app session information, and an opaque user identifier (u_…) · Destination country: United States · Timing and method: transmitted over TLS when a notification is sent, and, in mobile app version 2.0.8 and later, when the app terminates abnormally or reports an error and when the app starts (sessions) · Purpose: delivery of Android push notifications, and crash analysis and stability of the mobile apps · Retention and use period: (Cloud Messaging) the period needed to deliver the notification; (Crashlytics) crash records and identifiers are deleted automatically after 90 days · Method and effect of refusal: notifications can be refused by turning off notification permissions in device settings, in which case only notifications are not received; crash reporting cannot be turned off individually in the app, and users who do not want it can use the web (axit.team) instead of the mobile app (nothing is sent to Crashlytics from the web).
6-10 Paddle.com Market Limited (United Kingdom) — Payments
· Recipient and contact: Paddle.com Market Limited (privacy@paddle.com; 30 Old Bailey, London EC4M 7AU, United Kingdom) · Items transferred: the email address, payment method, and billing information that the user enters directly on the Paddle checkout, and the order information the Company sends with the payment request (workspace and user identifiers, plan, billing cycle, number of seats, and credit quantity) · Destination country: United Kingdom (and the processing locations disclosed in Paddle's privacy policy) · Timing and method: transmitted through the Paddle checkout and API (TLS-encrypted) when the user pays, changes a subscription, or requests a refund on the AXIT website · Purpose: order, payment, and tax processing as the merchant of record, subscription management, and refunds · Retention and use period: the period set by Paddle's privacy policy and by applicable laws such as tax and accounting laws · Method and effect of refusal: nothing is transferred unless the user pays on the web; refusal only means that paid purchases on the AXIT website cannot be used.
6-11 Mailgun Technologies, Inc. (United States) — Email
· Recipient and contact: Mailgun Technologies, Inc. (Sinch group, dpo@sinch.com) · Items transferred: the recipient addresses and content of emails the Service sends (sign-in verification codes; invitations containing the inviter's name and profile photo URL and the workspace name; notices about email changes and workspace deletion; and email replies sent by the AI), and emails arriving at service addresses (sender, recipients, subject, body, attachments, and headers) · Destination country: United States · Timing and method: transmitted via API (TLS-encrypted) when email is sent or received · Purpose: sending and receiving service email · Retention and use period: processed for the outsourcing period; sending and receiving logs are kept and then deleted under Mailgun's policies · Method and effect of refusal: emails necessary to provide the Service, such as sign-in verification codes, are difficult to refuse individually (no verification-code emails are sent when Sign in with Google or Apple is used); users who wish to refuse may stop using the Service.
6-12 Eleven Labs Inc. (United States) — Note Transcription
· Recipient and contact: Eleven Labs Inc. (dpo@elevenlabs.io; 169 Madison Ave #2484, New York, NY 10016, USA) · Items transferred: note recording audio files and the language setting · Destination country: United States · Timing and method: transmitted via API (TLS-encrypted) in workspaces designated by the Company (currently internal to the Company) when a user requests transcription of a recording or when automatic transcription runs after a recording is uploaded from the mobile app · Purpose: transcription and speaker separation of recordings · Retention and use period: processed for the period needed for transcription, as set out in ElevenLabs' privacy policy and service agreement · Method and effect of refusal: nothing is transmitted unless transcription is requested (recordings from the mobile app are transcribed automatically after upload); refusal only means that transcription (speaker separation) cannot be used.
7. Destruction Procedures and Methods
The Company destroys personal information for which grounds for destruction have arisen as follows. Destruction resulting from a user's withdrawal or deletion is carried out automatically by the Service; other destruction goes through an approval process by the department responsible for personal information protection. · Membership withdrawal: AXIT deletes the account information (email address, name, profile photo, and sign-in methods), device tokens for notifications, and private notes (including their recordings and transcripts). That person's "Chat with AXIT" and "Message yourself" conversations can no longer be opened by anyone, and AXIT deletes the files in them and any uploads that were never sent (including those in drafts and scheduled messages) and destroys the conversation records within 5 days. Conversations between people are also destroyed within 5 days once everyone in them has withdrawn. Google connections are disconnected under Section 15-8. Messages and files posted in channels and in conversations with other members, shared notes, and library materials remain as workspace content (Article 13 of the Terms of Service), and AI subscriptions, API keys, and external service integrations connected to the workspace remain as workspace settings until owners or admins disconnect them or the workspace is deleted. In records kept as required by law (payment, usage, credit, and customer inquiry records), the user is replaced with an irreversible "deleted user" marker. · Workspace deletion: after deletion is requested, the workspace can still be used as usual and the deletion can be canceled during a grace period of 7 days. When deletion is confirmed after the grace period, the workspace can no longer be opened by anyone; AXIT deletes the files uploaded to it (channel attachments, task outputs, library, note recordings, published pages, etc.) and its working folders, and disconnects Google connections under Section 15-8. The remaining data, such as messages, task records, and settings, is destroyed 3 days after the deletion is confirmed (within 5 days at the latest), while payment and usage records (including the workspace name attached to them), audit logs, and customer inquiries are kept for the periods in Section 3. A workspace in which no active member other than guests remains because the last member withdrew is deleted in the same way, without the 7-day grace period or the 3-day wait (except while it has a paid subscription). · Deleted content: when a message (including an AXIT result post) is deleted, AXIT deletes its content and its attachments and result files, and, if no one else has replied, its thread as well. AXIT also deletes attachments removed by editing, and deletes uploads that were never used in any message, draft, or scheduled message once 24 hours have passed since upload. However, a file that another live message still uses is deleted only after that message is also deleted. Files deleted from the library are erased 24 hours later, together with their previous versions, text, and summaries, and cannot be restored after that. When a channel is deleted, it disappears from the channel list, but the messages and files in it are not deleted at that time; they are destroyed together with the workspace when it is deleted. · Destruction methods: electronic files are permanently deleted using methods that make recovery or restoration impossible, and printed materials are shredded or incinerated. Information subject to statutory retention obligations is kept and used only to the extent necessary for the purpose set by the relevant law and is destroyed by the same methods once the retention period has elapsed. Backups and disaster-recovery copies are destroyed automatically once the retention periods set out in Section 3 have elapsed.
8. Rights of Data Subjects and How to Exercise Them
Users may at any time request access to, correction of, deletion of, or restriction of processing of their personal information, and may withdraw consent (including membership withdrawal). Requests may be submitted via in-service settings, or in writing, by email (contact@uniquegood.biz), or by phone (070-8706-1010), and may also be exercised through a legal representative or an authorized agent who submits a power of attorney. The Company will act on a request within 10 days of receipt and, if it refuses the request, will notify the user of the reasons and how to appeal. The Company does not collect personal information of children under the age of 14; if it becomes aware that such information has been collected, it will be deleted without undue delay.
9. Cookies and Other Automated Collection Devices; Behavioral Information
The Company uses cookies to maintain the Service (login sessions) and for usage statistics. Through the web analytics tool Google Analytics (Google LLC, United States), cookie identifiers, visit and usage behavior, and device information are collected for statistical and service-improvement purposes. Users may refuse the storage of cookies in their browser settings or install the Google Analytics opt-out browser add-on to refuse collection. However, refusing essential cookies such as login cookies may limit use of the Service. The Company does not collect behavioral information for personalized advertising purposes.
10. Security Measures
The Company applies the following measures. (Administrative) Establishing and implementing an internal management plan; minimizing access privileges and conducting periodic reviews. (Technical) One-way hashed storage of passwords; AES-256-GCM encryption of connected credentials and integration tokens; encryption in transit (TLS); encryption at rest of databases, file storage, search storage, and disaster-recovery copies; access controls; retention of access and audit records (within the Service they can only be viewed, and viewing is restricted); daily automated backups of the database and file storage and disaster-recovery copies kept in another region (Tokyo, Japan); and removal of personal information and secrets before error reports are sent. (Physical) Physical access control at the AWS data centers that run the servers.
11. Department Responsible for Personal Information Protection
For inquiries, grievance handling, and requests for access and other rights regarding personal information, please contact: Department in charge of personal information protection: Operations Team, Uniquegood Company Inc. / Phone: 070-8706-1010 / Email: contact@uniquegood.biz (7F, Merry Here, 8, Yeonmujang 13-gil, Seongdong-gu, Seoul, Republic of Korea).
12. Remedies for Infringement of Rights
Reports and consultations regarding personal information infringement may be directed to the following bodies: Personal Information Dispute Mediation Committee (Korea) (1833-6972, www.kopico.go.kr), Personal Information Infringement Report Center (dial 118 without an area code, privacy.kisa.or.kr), Supreme Prosecutors' Office (Korea) (1301, www.spo.go.kr), and Korean National Police Agency (182, ecrm.police.go.kr). If you disagree with the Company's decision on a request for access or other rights, you may file an administrative appeal under the Administrative Appeals Act (Korea).
13. Notice to Users Outside Korea
The Company processes personal information on the legal bases of performance of a contract, legitimate interests, and user consent. The Company does not sell or share personal information (including sharing for personalized advertising purposes). Users residing outside Korea may exercise the rights granted under the laws of their place of residence (access, deletion, restriction of processing, etc.) through the contact details in Section 11 above, and the Company will respond within the time limits prescribed by applicable law. You also have the right to lodge a complaint with your local supervisory authority.
14. Changes to This Policy
When this Policy is changed, notice will be given within the Service at least 7 days before the effective date; changes material to the rights of data subjects, such as changes to the items collected or the purposes of processing, will be announced at least 30 days before the effective date and notified individually by email or similar means. When this Policy is revised, a table comparing the provisions before and after the change is published together with this Policy, and previous versions are available upon request to the Company.
15. Google User Data
This Section explains how AXIT (the web at axit.team and the AXIT desktop and mobile apps) accesses, uses, stores, shares, retains, and deletes information received through Google APIs ("Google user data"). Where this Section differs from other Sections regarding Google user data, this Section prevails. AXIT's use and transfer to any other app of information received from Google APIs will adhere to Google API Services User Data Policy (https://developers.google.com/terms/api-services-user-data-policy), including the Limited Use requirements. Data received from Google Workspace APIs is also subject to the Google Workspace API User Data and Developer Policy (https://developers.google.com/workspace/workspace-api-user-data-developer-policy). The Company uses Google user data only to provide and improve the user-facing features described in this Section.
15-1 Sign in with Google (openid, email, profile)
· Information received: email address, name, and whether the email address is verified · Purpose: used only to create or find the AXIT account and sign the user in. · Storage: the email address and name are kept as account information and deleted upon membership withdrawal. The Google access token received during sign-in is used only once to confirm this information and is not stored, and the Google account identifier and profile photo are not stored.
15-2 Google Drive File Picker (drive.file, plus openid and email to identify the connected account)
· Scope of access: AXIT accesses only files the user picks with the Google file picker and files AXIT created, and does not view or search other files in Drive. · Connection: each member who is not a guest connects their own Google account. · Purpose: AXIT shows the name, type, and link of a picked file in the channel where it was shared, and fetches the file from Google each time an AI task requested by that user, or by another person in that channel (including guests in that channel), needs to read it. Tasks requested by other people can only read the file. Tasks requested by other people in other channels, tasks in other workspaces, and tasks started by email, webhooks, bots, or guests' posts cannot open the file. When the user saves a result to Drive or a task creates a document or sheet, AXIT creates a new file in the "AXIT" folder of the user's Drive, and later tasks can continue to edit that file. AXIT edits a file the user shared, shares a file, or moves a file to the trash only after obtaining the user's approval each time. · Storage: the refresh token is encrypted with AES-256-GCM and kept only on the server; it is not sent to the AI execution environment, and the server makes Drive calls on the task's behalf (while the file picker is open, only an access token that expires within one hour is passed to the user's browser). AXIT keeps the email address and identifier of the connected Google account, the ID of the "AXIT" folder, and records of shared files (file ID, name, type, link, and channel). File content read by a task is passed to the AI provider handling that task (Section 15-6), and files other than documents are downloaded to the channel's working folder for use and may remain in that working folder. File content left in the AI execution record (conversation log) is deleted within one hour after the task ends (after 3 days for tasks that stall). Task outputs remain as channel messages, and the "work process" view of a task keeps the commands the agent ran and the beginning of their output, which may include part of the file content. · Audit records: each Drive tool call is recorded in the audit log with the tool name and the file ID and name, without file content. Workspace owners and admins can view the audit log.
15-3 Google Drive Full Read Access and File Editing (drive.readonly; drive for file editing; plus openid and email to identify the connected account)
· How it is offered: this connection is offered through two Google apps. (a) The Company's internal app: in Company workspaces designated by the Company, only the Company's Google Workspace accounts can be connected. (b) The customer app: an external app for users outside the Company, to which personal Google accounts (Gmail addresses) can be connected as well as Google Workspace accounts. The customer app is in a testing stage before it completes Google's app verification (including the security assessment); for now it can be used only in workspaces designated by the Company, and while its publishing status in Google is "Testing", only accounts the Company has registered with Google as test users can connect it. · Who connects: in either app, members the Company permits (by default, workspace owners) connect their own Google account, and only the Google account with the same email address as their AXIT account can be connected. Guests cannot connect. · How to connect: before connecting, the app explains this and obtains the user's consent. When connecting through the customer app, after allowing access on Google's consent screen, the user confirms the Google account to connect once more on an AXIT screen before the connection is completed. While the customer app's publishing status is "Testing", under Google's policy the connection ends 7 days after it was made, and the user must connect again to keep using it (the settings screen shows the date it ends). · Information received: the list, names, types, and content of files in the connected account's Drive, and search results (with File editing turned on, also the location, owner (name and email address), sharing state, and edit permission of a file to be edited; the email addresses of the people a file is shared with are not kept, only counts by domain) · Purpose: only in tasks requested by the person who connected it (including scheduled tasks that person created themselves or approved when they were set up), the AI finds and reads the files it needs and shows the list of files read together with the result. This connection is not used for conversations or notes, for tasks started by email, webhooks, bots, or other people's posts, or for tasks requested by other people. New files are created only through the connection in Section 15-2, and this connection edits existing files only when File editing (below) is turned on. · Safeguards: the amount read per task is capped, and file content is passed to the AI as material separated from instructions. Tasks that read files through this connection run only in an isolated execution environment without shared file storage, with web search and external connections turned off. Sending the result of such a task outside AXIT (email replies, posts, GitHub, etc.) requires approval from all owners of the data, and some paths, such as published pages and saving to the library, are blocked. The work process of the task is visible only to the person who requested it, notifications hide the content, and the result content is masked when passed to later tasks requested by other people. · AI processing: sent only through the routes confirmed under Section 15-6 not to use the data for training. · Storage: the refresh token is encrypted with AES-256-GCM and kept only on the server; it is not sent to the AI execution environment. AXIT keeps a record of the files each task read (file ID, name, address, type, and number of characters read; search terms only as irreversible values) to show in the task view, and file content is not written to this record. When the task ends, its AI execution record is deleted immediately. Task outputs remain as channel messages. · File editing (optional feature, drive): in workspaces designated by the Company, the person who made this connection can separately turn on "File editing". Before it is turned on, the app explains this and obtains the user's consent, and the user must allow, once more and with the same Google account, the permission that Google's consent screen shows as "See, edit, create, and delete all of your Google Drive files" (drive). With the customer app, the user confirms the account once more on an AXIT screen, and while the publishing status is "Testing", the connection ends 7 days after this point. AXIT uses this permission only as follows: in tasks requested by the person who turned it on (including scheduled tasks that person created themselves or approved when they were set up), AXIT changes cell values and text in existing files that the connected account can edit: cell values in Google Sheets, Excel files (.xlsx), and CSV files, and text in Google Docs, Word files (.docx), PowerPoint files (.pptx), and text files (.txt, .md) (find and replace, and adding text at the end; for PowerPoint files, find and replace only). This includes files in My Drive, Shared with me, and shared drives, including files owned by other people. A file that is not a Google Doc or Sheet is downloaded, only the requested cells or text are changed, and the result is uploaded back to the same file in the same format (the file's name, location, and sharing stay the same). Before uploading, AXIT marks the version before the change to be kept permanently in the file's Google Drive version history, and if it cannot, it does not make the change. The user can download or delete that kept version in Google Drive's version management. AXIT does not change formulas or macros in Excel files: it does not edit cells that contain formulas, and it does not edit files that contain macros, password-protected files, or older-format Office files (.xls, .doc, .ppt). When it edits an Excel file, AXIT leaves the formulas unchanged and clears only their stored results, so that they are recalculated the next time the file is opened. Before each change, AXIT shows a confirmation card to the person who requested the task, shows only that person the file's name, location, owner, and sharing state and the content before and after each change, and makes the change only if that person allows it. Each change request is limited to what the confirmation card can show in full (50 places), and invisible formatting characters are refused or shown visibly on the card. The confirmation card visible in the channel does not include file names or the changes, and the changes are reviewed and allowed on the web and in the desktop app. If the person declines or does not answer within 10 minutes, nothing is changed. Files are not edited in tasks that include another person's Google Drive data. AXIT does not delete files, move them to the trash, change their sharing or permissions, move them to other folders, or transfer their ownership. The information received to show on the confirmation card (the file's location, owner, and sharing state, and the current content of the places to be changed) is not sent to the AI; it is stored encrypted and deleted once the card is answered or expires. Content downloaded to edit a file that is not a Google Doc or Sheet is likewise not sent to the AI and is not stored; it is discarded as soon as the edit is finished or is not made because it was declined or expired. Edited files are also kept in the record of files the task read. File editing can be turned off at any time on the same settings screen, and it stops immediately; confirmation cards still waiting for an answer are closed. Because the permission remains in the Google Account after File editing is turned off, the user can remove it by disconnecting Drive full read access or by revoking it directly in Google (Sections 15-8 and 15-9). If another AXIT connection uses the same Google account, disconnecting does not remove the permission on Google's side, so in that case the user revokes it directly in Google. File editing is turned off when this connection is made again or when the user's AXIT sign-in address changes.
15-4 Gmail (gmail.readonly, plus openid and email to identify the connected account)
· How it is offered: through the same two Google apps as in Section 15-3. (a) The Company's internal app: in the Company's workspace, only mailboxes on the Company's Google Workspace domain (uniquegood.biz) can be connected. (b) The customer app (the same app as the customer app in Section 15-3): personal Gmail mailboxes can be connected as well as Google Workspace mailboxes. The customer app is in a testing stage before it completes Google's app verification (including the security assessment); for now it can be used only in workspaces designated by the Company, and while its publishing status in Google is "Testing", only accounts the Company has registered with Google as test users can connect it. · Who connects: workspace owners and admins connect mailboxes. For a mailbox connected through the customer app, only the person who connected it can create or edit the receiving rules for that mailbox. · How to connect: when connecting through the customer app, the app explains this and obtains the user's consent before connecting, and after allowing access on Google's consent screen, the user confirms the Google account to connect once more on an AXIT screen before the connection is completed. While the customer app's publishing status is "Testing", under Google's policy the connection ends 7 days after it was made, and the user must connect again to keep using it. · Information received: when new mail arrives, to match it against the receiving rules, AXIT receives only its Gmail message and thread identifiers, labels, time received, and size, and the headers about where it was delivered (recipients, Cc, forwarding addresses, mailing list, and message ID). Only for mail matching the rules does AXIT receive the headers, including sender, subject, and date, and the body and attachments; mail that does not match is not stored. · Purpose: mail matching the rules is moved into posts in a designated public channel so that the people who can see that channel (including members of the workspace who have not joined the channel) can read it and use it in AI tasks, and if a rule includes an instruction, the AI processes the mail as instructed (for example, summarizing or classifying it) and replies in that post's thread. AI tasks started by mail are not given Drive, Calendar, or library tools. AXIT does not send mail from the connected mailbox, and does not modify or delete mail in it. · Storage: the refresh token is encrypted with AES-256-GCM and kept only on the server; it is not sent to the AI execution environment. Mail moved to a channel (sender, recipients, subject, body, and attachments) is kept as content of that channel until it is deleted under Section 7, and delivery records (such as the Gmail message ID) are kept so that the same mail is not moved twice.
15-5 Google Calendar (calendar.readonly, calendar.events)
· Connection: a workspace owner or admin connects one Google account to the workspace. The connection ends when the person who connected it leaves the workspace (removal or deactivation as a member) or deletes their AXIT account, and the workspace's calendar features stop until an owner or admin connects a calendar again (Section 15-8). · Information received: event information (title, time, description, location, attendees, etc.) in the calendars the connected account can see · Purpose: AXIT shows the events of the connected account's primary calendar to members who are not guests in "Today's schedule" in Notes and lets them create meeting notes from events. AI tasks requested by members who are not guests (including scheduled and automated tasks those members set up) can look up events and, as instructed, create, edit, or delete events; tasks started by email, webhooks, bots, or guests' posts, and tasks requested by guests, are not given calendar access. · Who can see it: events in the connected calendar are visible to all members of the workspace who are not guests and are used in AI tasks those members request. · Storage: the refresh token is encrypted with AES-256-GCM and kept only on the server; only an access token that expires within one hour is placed in the AI execution environment, and the AI task calls the Google Calendar API directly with that token. "Today's schedule" is kept in server memory for at most one minute for display. Event content read by an AI task remains in that task's AI execution record and is deleted within one hour after the task ends (after 3 days for tasks that stall). Notes created from an event keep the event ID, title, and start and end times until the note is deleted. Event information included in task outputs remains in channel messages and is visible to people who can view that channel.
15-6 AI Processing, Where Data Goes, and No Training or Advertising
· The Company does not use Google user data, or data derived from it, to develop, improve, or train generalized (not specific to a particular user or organization) AI or machine learning models, and does not provide it to third parties for that purpose. · AI processing: to perform tasks requested by the user or by members of the same workspace, Google user data may be passed to AI providers (Section 6) on the following basis. (a) Data read through Drive full read access (Section 15-3) and its results are sent only through routes the Company has confirmed do not use input data for model training. These routes are Claude and ChatGPT subscription accounts confirmed and recorded as having training turned off, Amazon Bedrock, which contractually does not use inputs for training, and, where confirmed and recorded, the Anthropic and OpenAI APIs; the Z.ai (GLM), Moonshot AI (Kimi), xAI (Grok), and Google (Gemini) engines are not used. Drive full read access is available only in workspaces where the Company has turned on "Google data protection", and its results cannot be saved to the library. (b) Other Google user data (files shared through the Drive file picker, calendar events, and mail moved from Gmail into channels, including its attachments) is also passed only to AI tasks that run through the same routes as in (a). It is not passed to tasks that run on other engines; the task that received it and later tasks in the same channel also run only through the routes in (a), and these tasks do not save their results to the library. · Other destinations: Google user data and results derived from it leave AXIT only to AWS as the service operation processor (servers, databases, file storage, and search in the Seoul Region, and the Tokyo disaster-recovery copies in Section 6-6); to Apple (APNs) and Google (Firebase Cloud Messaging), through which task result notifications pass (notifications hide the content of Drive full read access results); where sent outside at the user's instruction (email replies, or exports to GitHub, Notion, etc. connected by the workspace); and for security purposes, to comply with law, or in a merger, acquisition, or sale of assets with the user's explicit prior consent. Task outputs and mail posts in channels are indexed on the search servers in the AWS Seoul Region and converted into search vectors in Amazon Bedrock (Seoul Region). · No sale or advertising: the Company does not sell Google user data and does not use or provide it for advertising (including personalized, interest-based, or retargeted advertising), for data brokering or reselling, or for determining creditworthiness or lending.
15-7 Human Access
· Company staff do not read Google user data, except (1) where the user has explicitly agreed to staff viewing specific data (for example, in a support request), (2) where necessary for security purposes such as investigating bugs or abuse, (3) where necessary to comply with law, or (4) where the data is aggregated and anonymized for internal operations. The operational database query tool masks Drive full read access tokens and the content marked as derived from them. · Owners and admins of the same workspace can view that workspace's audit log and connection status as a feature of the Service.
15-8 Retention, Disconnection, and Deletion
· Disconnection: the person who made a connection, whatever their role, can disconnect it in "My settings › Account & security › Connected Google accounts", and owners and admins can also disconnect connections in "Workspace admin › Integrations" (Drive and calendar) and "Workspace admin › Receive by email" (Gmail) (Section 15-9). On disconnection, AXIT immediately deletes the stored token and, once it has confirmed that no other AXIT connection uses the same Google account, asks Google to revoke access (if Google does not respond, AXIT asks again shortly afterwards). Where AXIT does not ask Google to revoke access because another connection uses the same account or the connected account cannot be identified, or where the request does not reach Google, the user can revoke access directly in Google by the method in Section 15-9. After disconnection, no new data is received through that connection, and task outputs and mail posts already in channels remain until the user deletes them. When a Drive file picker connection is disconnected, the files that person shared can no longer be opened by any task, and the records of shared files remain with a revoked mark. Calendar and Gmail mailbox connections are used by the whole workspace, so disconnecting one stops the features that use it in that workspace; the mailbox address, Google account identifier, and receiving rules remain in the workspace settings. · Removal or deactivation of a workspace member: AXIT disconnects that person's Drive connections in that workspace (file picker and full read access) and the calendar and Gmail mailboxes that person connected there, in the same way as a disconnection (deleting the tokens and asking Google to revoke access under the same conditions). Connections in other workspaces are not affected. · Membership withdrawal: AXIT deletes that person's Drive connections (file picker and full read access) and tokens, records of shared files, and records of files read, replaces the body and work process of Drive full read access results with a deletion notice, and deletes the result files in the working folder. It deletes the calendar connections that person made and deletes the tokens of the Gmail mailboxes that person connected, which stops receiving mail from them (the mailbox address and receiving rules remain in the workspace settings). For all of these connections (Drive file picker and full read access, calendar, and Gmail), AXIT asks Google to revoke access under the same conditions as a disconnection. · Workspace deletion: when deletion is confirmed after the grace period (7 days), AXIT deletes the workspace's Google connections (Drive file picker and full read access, calendar, and Gmail) together with their tokens, asks Google to revoke access under the same conditions as a disconnection, and deletes Drive full read access results and the records of files read. · Backups: deleted data remains in database backups (Seoul, 7 days) and their Tokyo copies (7 days), in file storage backups (Seoul, 35 days), and in automated search server snapshots (Seoul, 14 days) until those periods elapse, and is then deleted automatically. The Tokyo replica of the file storage follows deletions in the original, and deleted uploaded files are also deleted from the upload storage (Amazon S3), where their previous versions are deleted after 30 days in Seoul and after 14 days in the Tokyo copy (Section 3). · The mobile apps keep some recent messages encrypted in the device's secure storage for at most 7 days. · Retention by AI providers follows each provider's policy in Section 6.
15-9 How to Revoke Access
· In AXIT: the person who made a connection, whatever their role, can see and disconnect the Drive, calendar, and Gmail connections they made in "My settings › Account & security › Connected Google accounts" on the web and in the desktop app. The mobile apps do not have this screen yet; mobile app users can disconnect on the web. Owners and admins can also disconnect connections in "Workspace admin › Integrations" (Drive and calendar) and "Workspace admin › Receive by email" (Gmail). What happens on disconnection is described in Section 15-8. If disconnecting on screen is not possible, for example because the person can no longer sign in, disconnection can be requested by writing to contact@uniquegood.biz. · In Google: access can be revoked at any time on the third-party connections page of the Google Account (https://myaccount.google.com/connections). After revocation, AXIT can no longer receive data through that connection. · Deletion requests: access to and deletion of stored Google user data can be requested by the methods in Section 8.
15-10 Security and Changes
· Google connection tokens are stored encrypted with AES-256-GCM, data in transit is encrypted with TLS, and databases, file storage, search storage, and backups are encrypted at rest. Permissions are checked on the server, and Drive tool calls are recorded in the audit log. · Before using Google user data in a way or for a purpose not described in this Section, the Company will first revise and announce this Policy and obtain the user's consent where required.
16. Revision History
Main changes in this revision, announced on October 29, 2026 and effective November 28, 2026 (previous version: effective October 23, 2026): (1) Sections 15-3 (Google Drive full read access) and 15-4 (Gmail): these connections can now also be made through a customer Google app (in a testing stage) in addition to the Company's internal app, so the Policy now describes how the two apps are offered, which accounts can be connected (including personal Google accounts), who connects, and how to connect (the explanation and consent before connecting, confirming the Google account to connect, and the 7-day connection limit while the publishing status is "Testing"), and states that only the person who connected a mailbox through the customer app can create or edit its receiving rules. (2) Section 15-3: the Policy now states that the connection is also used in scheduled tasks that the person who connected it created themselves or approved when they were set up. (3) Section 15-4: the information received to match mail against the receiving rules was corrected to match actual processing (the headers about where the mail was delivered, labels, etc.), and the Policy now states that moved mail becomes a post in a public channel that members of the workspace who have not joined the channel can also see, that the AI processes mail as a rule instructs and replies in the thread, and that AXIT does not send mail from the connected mailbox or modify or delete mail in it. (4) Section 15-3: the optional File editing feature (drive) was added, describing how it is turned on (the explanation and consent in the app, and allowing it once more with the same Google account), what it edits (cell values in existing Google Sheets, Excel, and CSV files and text in existing Google Docs, Word, PowerPoint, and text files, with the version before each edit of a non-Google file kept permanently in Google Drive's version history), that each edit is shown only to the person who requested the task and made only if that person allows it, what it does not do (deleting files, moving them to the trash, changing sharing or permissions, moving them between folders, transferring ownership, or changing formulas or macros in Excel files), and how to turn it off and remove the permission on Google's side. The table comparing the provisions before and after the change is available at https://axit.team/privacy?compare=2026-10-23, and the full text of the previous version is available at https://axit.team/privacy?version=2026-10-23 and also upon request to the Company.
Main changes in the revision announced on October 17, 2026 and effective October 23, 2026 (previous version: effective July 20, 2026): (1) Sections 5 and 6: Amazon Web Services (storage in the Seoul Region, disaster-recovery copies in the Tokyo Region, Japan, and Amazon Bedrock), Functional Software (Sentry), Google (Firebase Crashlytics and Firebase Cloud Messaging), Paddle, Mailgun, and ElevenLabs were added as processors and overseas transfers, and the items and conditions of transfers to Cloudflare, Apple, and AI providers were corrected to match actual processing. (2) Sections 1 and 2: the purposes and items now include sign-in methods (email verification code, Google, and Apple), mobile device tokens, error and crash report information, note recordings, received emails, Slack imports, and Paddle payment information. (3) Sections 3 and 7: the backup retention period of "7 days" was corrected to 7 days for the database, 35 days for file storage, 14 days for search server snapshots, and the retention periods of the Tokyo copies; the Policy now states that deleted files are also deleted from the upload storage (Amazon S3) and that their previous versions are deleted after 30 days in Seoul and 14 days in Tokyo, what is deleted and what remains on membership withdrawal, workspace deletion, and content deletion (including when deleted workspaces and withdrawn users' conversations are destroyed), the retention periods of audit log entries that are contract records (5 years) and of original copies of received emails (1 year), and how records subject to statutory retention are kept. (4) Section 4: a notice about information sent to external services connected by the user was added. (5) Section 6: the description of the basis for transfers for web analytics (Google Analytics) was corrected. (6) Section 10: encryption of integration tokens, encryption at rest, file storage backups and disaster-recovery copies, and the removal of information before error reports are sent were added, and the description of audit records ("protection against forgery and alteration") was corrected to the measures actually taken (within the Service they can only be viewed, and viewing is restricted). (7) Section 15 (Google User Data) was added. The full text of the previous version is available at https://axit.team/privacy?version=2026-07-20 and also upon request to the Company.
The Korean text is the authoritative version of this Policy. This translation is provided for convenience; the Korean version prevails in case of any conflict. This Policy was announced on October 29, 2026 and applies from November 28, 2026.