Privacy Policy
Last updated: 2026-11-16
NEXUPIRA (operated by Lofty Mems Co., Ltd., Tax ID 54335019; "we", "us", or "the Service") is an AI-powered integration platform that lets users connect their productivity tools (Google Workspace, Slack, GitHub, Notion, Linear, Facebook, Instagram, Threads, and others) and automate workflows through natural-language instructions. This Privacy Policy describes what we collect, how we use it, and what control you have.
1. Information We Collect
When you use NEXUPIRA, we collect:
- Account information: email address, name, and (if you register with a password) a salted password hash.
- OAuth tokens: access tokens and refresh tokens from services you choose to connect. These are encrypted at rest with AES-256-GCM and used only to call APIs on your behalf.
- Content you upload: documents, images, and files you attach to chats. The contents of uploaded documents are encrypted at rest (see section 5).
- Chat messages and workflow definitions: stored to let you resume sessions and re-run automations.
- Operational telemetry: server error logs and performance metrics, and page-load and error reports from the app in your browser (used to diagnose bugs and keep the Service reliable). We do not record your screen or sessions.
- Support requests and bug reports: what you write to us or report from the app, and your contact details so we can reply.
- Website and usage analytics: page and conversion statistics collected through Google Analytics; see "Website analytics" in section 6.
Connected customer messaging channels
If you connect LINE, WhatsApp Business, Telegram or Discord, we process messages and attachments received by that connection, channel and sender identifiers, customer display/profile names, short text excerpts of recent messages and delivery results. This processing starts when messages arrive, even if you have not enabled an inbox automation. If you enable the inbox, we also prepare reply drafts and follow-up lists. You choose which channels to connect; you are responsible for informing your own customers about how you use their messages.
When voice transcription is enabled, audio is downloaded and sent to the provider of the API key you use for transcription (for example OpenAI or Google). The transcript may be used to draft a reply. The standard customer-service inbox prepares drafts for your review and sends replies after approval. WhatsApp replies outside its customer-service window use an approved template only where the workflow and account configuration support it; a template is different from the drafted text. If you link an owner notification channel, we send alerts and links to that linked account; alerts can include a short excerpt of a reply draft and a daily summary.
You can disconnect a channel and stop its inbox automation from Connections and Automations. Already stored messages, drafts and run history remain subject to the data-retention and deletion controls described below. Each connected platform also processes messages under its own terms and privacy policy.
Contact data you add
You can add other people's contact details to the Service, for example business contacts you want to reach, and send them messages from your own connected mailbox. We process that contact data on your behalf and only to carry out your instructions. You are responsible for having a lawful basis to collect and contact the people you add, for telling them how you use their data where the law requires it, and for honouring their requests to stop or to have their data deleted.
2. Google User Data
When you connect a Google account, you grant NEXUPIRA access to a limited set of Google APIs. The scopes we request are:
calendar&calendar.events— read and manage your Google Calendar events when you ask an automation to do so.drive.file— create, read, and modify only the files that NEXUPIRA itself creates or that you explicitly open through NEXUPIRA. We never read your entire Drive.documents,spreadsheets,presentations— edit the Google Docs, Sheets, and Slides files that NEXUPIRA creates on your behalf.gmail.send— send emails from your account when your workflow asks us to.gmail.compose— create draft emails.gmail.labels— list your existing Gmail labels.
Limited Use disclosure
NEXUPIRA's use and transfer of information received from Google APIs adhere to the Google API Services User Data Policy, including the Limited Use requirements. Specifically:
- We use Google user data only to provide or improve user-facing features that are visible in the NEXUPIRA interface.
- We do not transfer Google user data to third parties except as necessary to provide those features, comply with law, or prevent fraud.
- We do not use Google user data to serve advertisements.
- We do not allow humans to read Google user data unless we have your explicit consent for specific messages, it is necessary for security purposes (e.g. investigating abuse), to comply with applicable law, or the data is aggregated and used for internal operations in line with these requirements.
3. Facebook, Instagram and Threads Data
If you connect a Facebook Page, an Instagram professional (Business or Creator) account or a Threads account, we receive data from Meta Platforms through the permissions you approve on Meta's consent screen. We use it only to provide the features you use in NEXUPIRA, in line with Meta's Platform Terms and Developer Policies.
What we receive
- Facebook Page: your Facebook user ID and name; the Pages you granted, and the name and ID of the Page you connect; the Page's posts (text, time, link, and reaction, comment and share counts); and comments on those posts, including the commenter's name and ID.
- Instagram: your username, account ID and account type; your posts (caption, media type, link, time, like and comment counts); comments and replies on your posts, including the commenter's username; post insights (reach, likes, comments, saves, shares, total interactions) and account insights (reach, accounts engaged, total interactions, follows and unfollows); and how many more posts you can publish through the API in the current 24 hours.
- Threads: your username and ID; your posts; replies and conversations under your posts; post and account insights (views, likes, replies, reposts, quotes, shares, followers); and, when you search, public Threads posts that match your keyword.
How we use it
- We read this data only when you ask the assistant in a chat, or when an automation you set up runs.
- The assistant can only prepare drafts. A post is published, a reply is sent or a comment is hidden only when you press Publish, Send or Hide on the draft card, or when you press Approve in the review step of an automation you set up (for example a weekly post), and each press sends it once. Nothing is posted or replied to without your approval. If you schedule a Facebook post, it is sent to Meta when you press Publish and appears at the time you chose.
- We do not like, follow, repost or delete anything on your accounts.
AI processing
When you ask the assistant about this data, the posts, comments or numbers it needs are sent to the AI model handling your request: the platform AI (Anthropic or Google Vertex AI, see section 6) or the provider of the API key you connected. We do not use data from Meta to train or evaluate AI models, and we do not sell it or share it with advertisers or data brokers.
Storage
- Access tokens are encrypted at rest with AES-256-GCM. For Facebook we keep only the Page access token, except that when you grant several Pages you can post on, each Page's token is kept encrypted for up to 15 minutes so you can choose which Page to connect, and then deleted; for Instagram and Threads we keep a long-lived token that is refreshed automatically while you stay connected. We also keep the connected account's ID and name (Page name or username), and for Facebook your Facebook user ID and name and the list of Pages you granted (names and IDs).
- This data is stored on our servers in Japan (AWS, Tokyo) and processed by the service providers listed in section 6: Amazon Web Services (Japan), Google Cloud Vertex AI (United States), Anthropic (United States) and Cloudflare (United States), plus the provider of your own API key if you use one.
- Posts, comments and numbers the assistant reads in a chat are saved in that chat's history so you can come back to it. They are kept until you delete the chat or your account. If the assistant's reply quotes them, a short note may also be saved in your memories, which you can review and delete on the Memories page.
Disconnecting and deleting
- Disconnect in NEXUPIRA: on the Connections page, disconnecting Facebook, Instagram or Threads deletes the stored token and connection details straight away. It does not remove NEXUPIRA's permission inside Meta; you can remove it in Facebook's "Business integrations", Instagram's "Apps and websites" or Threads' "Website permissions".
- Remove NEXUPIRA in Meta: when you remove NEXUPIRA from Facebook, Instagram or Threads, Meta notifies us, and we delete the token and connection details for that Meta account from every NEXUPIRA account it was connected to.
- Data deletion request: when you ask Meta to delete the data NEXUPIRA received, we delete the same token and connection details and give Meta a confirmation code. Meta shows you a link where you can check that the deletion is complete. We keep a record of each request (your Meta user ID, the confirmation code and the time) to show that we handled it.
- These steps do not delete chats that contain data from Meta. Delete those chats yourself, delete your account, or email support@nexupira.com and we will delete them for you.
4. How We Use the Data
- To execute the automations and chat instructions you give us.
- To send prompts and the context you select to the AI model provider handling the request: the platform AI provider, or the provider whose key you connected (see section 6).
- To answer your support requests and fix the problems you report.
- To protect the Service and the free trial from abuse (for example bot checks at sign-up and blocking repeat free-trial sign-ups).
- To debug crashes and improve reliability (error traces and logs).
- To authenticate you on subsequent sign-ins.
We do not sell your data, use it to train AI models, or share it with advertising networks.
Legal bases
- Performing our contract with you: providing the Service you signed up for.
- Legitimate interests: security, abuse prevention, improving the product, and measuring how people find and use the website.
- Legal obligations: keeping tax and invoice records.
Requests from public authorities
If a public authority asks us for users' personal data, we review whether the request is lawful, challenge it if we consider it unlawful, disclose only the minimum information necessary, and keep a record of each request, our response, the legal reasoning and the people involved.
5. Data Storage & Security
- Tenant application data is isolated in PostgreSQL schemas. Shared account and security records are held separately with server-side access controls.
- Credentials and connected-service tokens (such as OAuth tokens and your own API keys), and the contents of uploaded documents, are encrypted at rest with AES-256-GCM using keys managed by NEXUPIRA.
- Other data, such as chat messages, images attached to chats and the search index built from your documents, is protected by account isolation, access controls and encrypted transport, but is not separately encrypted by NEXUPIRA.
- Transport is HTTPS / TLS 1.2+ only.
- We apply the principle of least privilege: OAuth scopes are the minimum needed for the feature you invoked.
6. Third-Party Services
To deliver the Service we send necessary data to:
- AI model providers — for the purpose of generating the response you asked for. When you use platform AI, the request is processed by a model provider we configure: currently Google (Vertex AI) and Anthropic (Claude models, through the Anthropic API and/or Amazon Bedrock on AWS); which one handles a request depends on the active platform configuration. When you use your own API key, it goes to the provider you selected (for example OpenAI, Anthropic, Google Gemini, Groq, OpenRouter, or a custom endpoint).
- Your LLM API keys and OAuth tokens are stored encrypted (AES-256-GCM) and used only to serve your own requests. Prompts, attachments and retrieved documents are sent to the provider handling the request under that provider's own terms. We do not use customer content to train models, and we send your content only to the platform's configured provider or to a provider you connected yourself.
- Embedding providers — to build the search index for your documents, document text is sent to the embedding model we configure for the platform (such as Google Vertex AI), or to the provider you choose in Settings.
- Services you connect (Google, Slack, GitHub, Notion, Linear, LINE, Meta (Facebook, Instagram, Threads, WhatsApp), Telegram, Discord and others) — these are your own accounts. We send or fetch data there only when you, or an automation you set up, uses them, and each service handles that data under its own terms.
- Voice transcription — when voice processing is enabled, audio is sent to the provider of the API key you use for transcription, such as OpenAI or Google.
- Cloud hosting (Amazon Web Services) — our servers, databases and backups run on AWS.
- Web search (Tavily) — when a teammate searches the web for you, the search query is sent to Tavily, or to the search provider whose key you connected.
- Cloudflare — all traffic to our website and app passes through Cloudflare's network for content delivery and protection, including Turnstile bot checks on sign-up and sign-in. Cloudflare processes your IP address and the requests that pass through it.
- Amazon SES (AWS) — sends account and service emails, such as sign-up verification, password resets and notifications. It processes your email address and the content of those emails.
- Slack and Notion (internal operations and support tools) — when you send a support request or bug report, our Slack receives only a ticket number and a link; the report itself, including your contact details, is kept in Notion and in our internal admin tool.
- Payment processor — Paddle.com (as Merchant of Record) processes payments for paid plans; see section 10.
- Logs and monitoring — error logs, performance metrics and browser reports are stored in OpenObserve, which we host ourselves on our own servers; no third-party analytics company receives them. Logs are kept for 7 days.
We do not send your data anywhere you have not explicitly connected, except to our own operational infrastructure, the service providers listed in this section, and Google Analytics as described under "Website analytics" below.
International data transfers
Your data is processed mainly in Taiwan, Japan and the United States, by us and by the service providers listed in this section; some providers (for example Cloudflare's network) may process it in other countries. When data is transferred outside the country you are in, we rely on those providers' contractual data-protection safeguards.
Website analytics (Google Analytics)
nexupira.com uses Google Analytics 4, provided by Google LLC and loaded through Google Tag Manager, to understand how visitors find and use our website. It sets first-party cookies in your browser (_ga, _ga_*) and records the pages you view, where you came from (for example UTM parameters in a link), your device type and approximate region.
When you sign up for NEXUPIRA, finish onboarding, have an AI teammate complete a task for the first time, start the Pro trial, open a checkout or complete your first payment for a plan, our server sends one additional event to Google Analytics. The event contains only the event name, the random identifier from that cookie, how you signed up or started the trial (email, a third-party sign-in, or from your account), the name of the role or starter task you chose, and for a checkout or payment the plan, the amount paid excluding tax, the currency and a scrambled reference that cannot be turned back into the payment number. Because a payment is confirmed by our payment provider rather than by your browser, we keep that random identifier on our server for at most two days after you open a checkout, and delete it once the payment event is sent or your account is deleted. The event does not contain your email, name, account ID, conversations or files. If your browser has no such cookie (for example because you block Google Analytics), nothing is sent.
We use these statistics only to measure which channels bring people who actually use the product, never for advertising. For how Google processes this data, see the Google Privacy Policy.
Cookies
- Necessary: sign-in and security cookies (your session, and short-lived sign-in state while you connect a service), a few preference cookies, such as your language choice on nexupira.com, and cookies that record which invite link you came from and protect invites from abuse.
- Analytics: nexupira.com loads Google Analytics 4 through Google Tag Manager, which sets the
_gaand_ga_*cookies described above. - Sign-up source: the first time you open nexupira.com we set a first-party
nx_ftcookie (kept 90 days) holding only the name of the site that linked you (for example google.com, never the page or what you searched), the UTM tags in the link, whether an ad link was clicked, the first page you opened and when. If you create an account, it is stored with your account so we can tell which channels bring sign-ups; it is not shared with anyone. Browsers that send Global Privacy Control do not get it. - Monitoring: the app sets a
_oo_scookie that groups its page-load and error reports for one visit (see "Operational telemetry" in section 1); it contains no account details.
You can block or delete cookies in your browser settings, or stop analytics with the Google Analytics Opt-out Browser Add-on. If you block necessary cookies, you will not be able to sign in.
7. Your Choices
- Disconnect a service at any time from the Connections page. We delete the tokens we stored for it. Where revoking at the provider cannot cut off your other connections (currently GitHub and HubSpot), we also ask the provider to revoke our access. We do not do this for Google or Slack, because revoking there would also disconnect your other NEXUPIRA accounts that use the same Google sign-in, or your colleagues in the same Slack workspace; you can remove our access yourself in your Google Account or Slack settings. For Facebook, Instagram and Threads we delete the token but do not revoke our access at Meta; section 3 explains how to remove it there.
- Delete documents from the Documents page; deletion is immediate and permanent.
- Delete a meeting room: it moves to Recently deleted, where you can restore it for 7 days; after that it is permanently deleted automatically.
- Delete your account yourself in Settings → Account, or by emailing support@nexupira.com; we delete or anonymise your data within 30 days (see section 8 for what we keep).
- Request a copy of your data by writing to support@nexupira.com; we reply within 15 days. If we need more time, we may extend this once by up to 15 days and will tell you why. There is no self-serve export at this time.
- Turn off website analytics by blocking or clearing cookies in your browser, or by installing the Google Analytics Opt-out Browser Add-on.
You can also revoke our access directly from your Google Account settings at myaccount.google.com/permissions.
8. Data Retention
Chat messages, memories, documents and automations are kept until you delete them or your account. A meeting room you delete stays in Recently deleted for 7 days and is then permanently deleted; documents you delete are removed permanently straight away. OAuth tokens are kept until you disconnect the service or delete your account. Server logs are kept for 7 days, and records of background jobs for up to 3 days.
When your account is deleted, we delete or anonymise your data in our live systems within 30 days, and messages arriving on channels you had connected are no longer routed to it. Copies in our backups are removed as the backups are routinely replaced; until then, backups are used only to restore the Service after a failure. We keep only: records the law requires us to keep, such as invoices and payment records under tax law; billing and usage records; support requests you sent us; security audit records; an internal record of the deletion with your email address masked; records of data-deletion requests received from Meta (see section 3); and the one-way hash described below.
When you start the free trial, we record a one-way hash of your email address (normalised, so variants of the same address match). We keep it after your account is deleted, only to prevent repeat free-trial sign-ups. We do not keep the email address itself, and we use the hash for nothing else.
9. Children's Privacy
The Service is only for people aged 18 or older (see Terms of Service §3). We do not knowingly collect information from anyone under 18; if we learn that we have, we delete it.
10. Payment Data, PCI DSS & PDPA Notice
To process subscription, add-on, and custom transactions, the data collector is Lofty Mems Co., Ltd. (Tax ID 54335019). Notice under Article 8 of Taiwan's Personal Data Protection Act (PDPA):
- Payment data collected: billing contact name, billing email, (for business buyers, where entered at checkout) company name and tax ID, and transaction amounts/records.
- Card data & PCI DSS: Full card data (number, expiry, CVV) is collected and processed directly by Paddle.com, a PCI DSS Level 1 provider acting as our Merchant of Record. We never store full card numbers, and we do not process or transmit them on our own systems, so our environment falls under PCI DSS SAQ A scope; we keep only transaction and subscription identifiers for billing and renewals. Payment pages use TLS encryption.
- Legal basis: performance of our service contract with you (provisioning, renewals and billing records), compliance with legal obligations, and our legitimate interests in fraud prevention and reconciliation.
- Purposes: billing and subscription management, customer service, contract performance, and legal compliance.
- Retention: until the purpose ends or statutory accounting/tax retention periods expire.
- Use region/recipients/method: mainly in Taiwan, Japan and the United States (see "International data transfers" in section 6), used electronically; recipients include Paddle.com (our Merchant of Record), our cloud provider (AWS), and lawful authorities.
- Your rights: access, review, copies, correction, and to stop collection/processing/use and delete (subject to law). For international users we also honour GDPR/CCPA-style rights, including data portability (we provide a copy of your data on request), restriction of processing, and the right to object.
- Effect of non-provision: without required payment/billing data, transactions and paid services cannot be provided.
- Contact to exercise rights: support@nexupira.com. We reply to access and copy requests within 15 days; if we need more time, we may extend this once by up to 15 days and will tell you why.
11. Changes to this Policy
We will update the "Last updated" date above whenever this policy changes. Material changes will be announced by email or in-product notice before they take effect.
12. Contact
General privacy, data-deletion requests, or to exercise your rights: support@nexupira.com. Account security and unauthorised-access reports: support@nexupira.com.
The company responsible for your personal data: Lofty Mems Co., Ltd. (嶸惠有限公司), No. 75, Xing'er St., Taoyuan Dist., Taoyuan City 330074, Taiwan.