Benzaiflow logo
Language
Back to home

Privacy Policy

Your data, handled with care.

This policy explains how Monocursive collects, uses, shares, and protects personal data when you use the Benzaiflow app, website, waitlist, and related services.

Effective and last updated: November 10, 2026

1. Data controller

Benzaiflow is operated by Monocursive, a company based in France.

For personal data processed through the Benzaiflow app, website, waitlist, support channels, and account systems, Monocursive acts as the data controller.

Privacy contact: contact@monocursive.com.

2. Data we collect

Depending on how you use Benzaiflow, we may collect or process the following categories of data:

  • Account and authentication data: name, email address, account identifiers, session identifiers, Sign in with Apple or Google sign-in metadata, anonymous guest account identifiers, and, for AI assistants you connect, the assistant's name, the access you granted, access tokens (stored as hashes), and when they were last used.
  • User content: tasks, projects, notes, chat messages, onboarding answers, profile preferences, work schedule settings, reminder settings, life markers, season summaries you choose to share, reports about Noji's messages, and feedback or support messages you send us.
  • Calendar and reminder data: titles, descriptions, times, timezones, locations, notification settings, task links, and status for entries you create in Benzaiflow's first-party calendar.
  • Subscription data: entitlement status, purchase product identifiers, renewal or expiration status, RevenueCat customer identifiers, and store transaction metadata needed to unlock paid features and prevent abuse.
  • Device and notification data: platform, app version, locale, timezone, device push tokens, notification permission status, and delivery state for reminders or app nudges.
  • AI input and output data: task text, chat context, profile facts, calendar context, photos you choose to scan, and generated responses or task suggestions needed to provide AI-assisted planning features.
  • Consent records: your answer to the “Use cloud AI?” question, the version of its wording you answered, and when you were asked and when you answered.
  • Technical and security data: IP address, request metadata, server logs, error logs, rate-limit records, fraud-prevention signals, and diagnostic information needed to run and secure the service.
  • Website, analytics, attribution, and waitlist data: email address, optional name, preferred platform, waitlist answers, language preference, pages viewed, referral information, browser/device categories, engagement duration, scroll depth, outbound-link destinations, first-party campaign and store-click events, and coarse country information.

3. Why we use this data

We use personal data only for purposes connected to operating, securing, improving, and supporting Benzaiflow:

  • To create and maintain accounts, including guest accounts and account linking.
  • To save, organize, sync, and display your tasks, projects, routines, reminders, calendar context, chats, and preferences.
  • To provide AI-assisted features such as task decomposition, task enrichment, daily planning, chat support, memory review, profile personalization, and reading photos you choose to scan.
  • To create the share links you ask for, show them to the people you send them to, and let AI assistants you connect act on your account within the access you grant.
  • To receive and review the reports you send us about messages Noji wrote.
  • To send reminders, push notifications, app nudges, verification emails, support replies, and service messages.
  • To manage subscriptions, restore purchases, apply Plus entitlements, process webhooks, and prevent subscription abuse.
  • To measure aggregate website use and understand which first-party campaigns lead to site visits and app-store clicks.
  • To secure the service, detect misuse, enforce rate limits and AI quotas, debug errors, monitor reliability, and maintain audit logs.
  • To comply with legal obligations and respond to lawful requests.

4. GDPR legal bases

Where GDPR applies, we rely on the following legal bases:

  • Contract or pre-contract steps (Article 6(1)(b) GDPR): to provide the app and the features you ask for, manage accounts, operate paid features, create the share links you request, connect the assistants you authorize, write the steps you ask for on the website's “Make it smaller” page, process waitlist requests, and respond to service requests.
  • Legitimate interests (Article 6(1)(f) GDPR): to secure Benzaiflow, prevent abuse, debug issues, improve reliability, understand feature and website performance, measure aggregate acquisition through Plausible and our limited first-party attribution, and maintain necessary business records. We limit these measurements to what is proportionate and do not use them for targeted advertising.
  • Consent (Article 6(1)(a) GDPR): for cloud AI in Synced, that is sending your content to Anthropic for chat, task breakdowns and details, plan suggestions, photo scans, and the other cloud AI features described in section 5; and for optional permissions or communications where consent is required, such as push notifications and optional marketing. You can withdraw consent at any time: for cloud AI, turn off Settings → Account & data → Cloud AI. Withdrawing does not affect processing that took place before.
  • Legal obligation: where processing or retention is required by applicable law.

5. AI-assisted features

Benzaiflow uses AI to help break down tasks, summarize context, suggest plans, enrich task metadata, and respond in chat.

Anthropic provides all of Benzaiflow's AI in the cloud: chat and onboarding, task breakdowns and enrichment, summaries, titles, profile generation, and planning-related structured outputs. Relevant messages, onboarding answers, task context, profile facts, calendar context, and app state may be sent to Anthropic when needed for the feature you use, and in Synced only once you have allowed it (see the next paragraph).

In Synced, the app asks “Use cloud AI?” before any of your content is sent to Anthropic, and lists what Anthropic would receive: what you write to Noji, the text of your tasks and captures, photos you choose to scan, your calendar events and reminders, and the profile notes kept about how you work. Nothing from your account is sent to Anthropic until you choose “Allow”: Benzaiflow's servers check your answer before calling Anthropic for your account. If you choose “Not now”, the app keeps working without AI answers, and your tasks, plan, timer, and calendar stay available. You can change your answer at any time with the switch at Settings → Account & data → Cloud AI; turning it off stops new requests to Anthropic and does not delete what you have already saved in the app. If we change what is sent, we will ask you again. Versions of the app from before this question existed (before 1.11.0) do not ask it: an account that has only ever been opened with such a version keeps that version's cloud AI behaviour until a current version of the app opens it and asks.

In our production service, requests to Anthropic pass through Cloudflare's AI Gateway, which runs on Benzaiflow's Cloudflare account. The gateway can keep logs of these requests and of Anthropic's answers, including the content sent for the feature you used, and can cache answers; those logs are kept by Cloudflare on Benzaiflow's account. Requests from the website's “Make it smaller” page ask the gateway not to log or cache them.

In Synced, you can choose a photo to scan for tasks, events, and reminders. The photo you pick is sent to Benzaiflow's servers and, through the AI Gateway described above, to Anthropic, which reads it; this is part of cloud AI and needs your consent. Benzaiflow's database keeps what was found in the photo and a fingerprint (a hash) used to recognise a repeated request, not the photo itself.

Voice notes are turned into text by your device's own speech recognition (Apple's on iOS, Google's on Android), which may process the audio on the device or on that provider's servers under its own terms. Benzaiflow receives only the resulting text, never the recording.

The “Make it smaller” page on our website (benzaiflow.com/shrink) works without an account. The text you type there is sent to Benzaiflow's servers and to Anthropic to write the steps. To limit abuse, the page may use Cloudflare Turnstile. A link made there, or shared from the app, holds the title and steps inside the link itself, in the part after the “#”, which browsers do not send to servers: opening such a link does not send its steps to Benzaiflow, and Benzaiflow keeps no copy of these links. The page removes that part from the address bar before analytics load, and the page's analytics events record only that steps were made or a link was opened, never their text. An image saved from the page is drawn in your browser and is not uploaded. Anyone you send such a link to can read the steps it holds.

The app runs in one of two modes, chosen when you start and changeable in Settings. In Synced, your content is stored in your account on our servers and the AI features above run in the cloud. In On this device, your tasks, chats, profile, and settings are stored only on your device, with no account: Benzaiflow does not receive them unless you send us a report about one of Noji's messages (see below), and AI, on devices that support it, runs on the device itself. Device backups you turn on (iCloud or Android backup) may include that data under the platform's terms.

On-device AI uses a model file (Gemma 4 E2B by Google DeepMind) that you can choose to download once from Hugging Face. The download goes directly from your device to Hugging Face, which receives your device's IP address and ordinary request metadata under its own privacy policy. No task, chat, or profile content is sent.

In On this device, crash reports are off unless you turn them on in Settings, and the app still checks Expo for app updates. Moving between modes happens only when you start it: moving to Synced uploads your content from the device to your account; moving to On this device copies it to the device, and your account's copy remains until you choose to delete it.

You can report something Noji wrote: a chat message or, from Settings, words you paste or describe. With a signed-in Synced account, the report is stored with your account's feedback: the reported words (at most the first 500 characters of a chat message, or 1,500 of what you paste), your note, for a chat message the identifiers of the message and conversation, and your device's platform, system version, and app version. In On this device, and from a guest account in Synced, sending a report sends the reported words, your note, the app's version, language, and platform, and which model wrote them (the on-device model's name when there is one, or “cloud” for a guest account) to Benzaiflow's servers, which forward them by email (through Resend) to our team; that report carries no name, account, or device identifier. Benzaiflow's report service uses the network address the report comes from only to limit how many reports can be sent, in short-lived rate-limit records and, for the daily limit, as a keyed hash; like any request, it is also processed by Cloudflare, which hosts our servers. Nothing is sent until you tap to send the report.

We do not sell your task, chat, calendar, or profile content. We do not use your personal content to build advertising profiles.

AI output can be incorrect or incomplete. You should review generated suggestions before relying on them, especially for calendar changes, reminders, or important commitments.

6. Optional permissions and account services

Some app features require optional permissions or third-party account services. You can use Benzaiflow without enabling every permission. Benzaiflow's calendar and reminders are first-party features and do not require linking an external planning account.

  • Push notifications: used to deliver reminders, timer updates, and app nudges if you grant notification permission.
  • Apple and Google sign-in: used for authentication and account linking.
  • Subscriptions: purchases are processed by Apple App Store or Google Play billing and entitlement state is managed through RevenueCat.
  • Photos: if you choose a photo to scan, the app asks for access to your photos where the system requires it, and sends only the photo you pick, as described in section 5.

7. Data sharing, processors, and no resale

We share data only with service providers needed to run Benzaiflow, provide user-requested features, process payments, or meet legal obligations.

AI assistants you connect to your account and people you give a share link to are not our processors: they receive data because you chose to share it with them (see section 8).

We may disclose data if required by law or to protect legal rights and platform security.

We do not sell, rent, trade, or resell your personal data.

Current processors, service providers, and platform recipients include the following. Where a provider acts as our processor, we require it to handle data only for the contracted service and with appropriate safeguards. App stores and identity providers may act as independent controllers for their own account or transaction processing under their privacy notices.

  • Cloudflare: website and API hosting, edge routing and security (including Turnstile checks on the website's “Make it smaller” page), Workers, Durable Objects, Workflows, AI Gateway (which passes requests to Anthropic and can log and cache them on Benzaiflow's Cloudflare account, as described in section 5), operational logs, and derivation of a coarse country code for first-party attribution.
  • OVHcloud: infrastructure hosting our PostgreSQL databases for account data, user content, service state, and first-party attribution events.
  • Anthropic: AI processing for all cloud AI features, including photos you choose to scan (in Synced, only after you allow cloud AI) and the website's “Make it smaller” page.
  • Hugging Face: hosting of the optional on-device AI model file, which your device downloads directly when you choose to; it receives your IP address and request metadata, not your content.
  • Google: optional Google sign-in, Google Play distribution and billing, and Android platform, push, or speech-recognition services where applicable.
  • Plausible Analytics: privacy-focused, aggregate website audience and performance measurement.
  • RevenueCat: subscription purchase validation, entitlement management, and related transaction state.
  • Resend: account verification, service, support, and other transactional email delivery, including the reports about Noji's messages sent from On this device.
  • Sentry: privacy-configured crash and error monitoring for the app and API; diagnostic payloads are restricted and scrubbed to reduce user-content and credential exposure.
  • Expo: app delivery services and push-notification delivery to the relevant Apple or Google push service.
  • Apple: optional Sign in with Apple, App Store and TestFlight distribution, in-app purchases, and Apple Push Notification service.

8. Links you share and assistants you connect

Two features available in Synced let you share your data yourself, with people or tools you choose.

Patterns & reports: you can create a link that shows a summary of a season of what you tracked in the app, for example to hand to a doctor or a coach. You choose the sections it shows; they can include the season described in words by AI and your life markers, which are off by default. The link contains a long random code, of which we store only a hash. The page shows a copy made when you create the link: later changes in the app do not update it. Anyone who has the link can open the page, without signing in, until the link expires 30 days after you create it or until you turn it off in Patterns & reports; we record when it was last opened. The page asks browsers and search engines not to store or index it. Someone who has opened it may have kept their own copy.

Connected assistants: with a signed-in account, you can connect an AI assistant that supports the Model Context Protocol (MCP) to your Benzaiflow account, by approving it on Benzaiflow's web app or by creating a personal access key. The assistant can then read and change your data within the access you grant: capturing tasks, reading your plan and tasks, completing, deferring, scheduling, and breaking down tasks, focus sessions, projects, saved memories, and your profile if the access you grant includes it. Actions that use Benzaiflow's cloud AI, such as breaking a task down, follow your answer to “Use cloud AI?”. An approved assistant stays connected as long as it keeps renewing its access, which lapses after 30 days without use, and an access key expires after 90 days by default and after one year at most; you can disconnect an assistant or revoke a key at any time in Settings → Account & data → Connections. What an assistant, and the company that provides it, does with the data it reads is governed by that provider's own terms and privacy policy, not by this policy.

9. Data retention

We keep personal data only as long as needed for the purposes described in this policy, unless a longer period is required or permitted by law.

  • Account data is kept while your account is active and for a limited period afterward if needed for security, legal, billing, or dispute reasons.
  • Tasks, projects, profile data, chats, routines, reminders, calendar entries, and service state are kept until you delete them, delete your account, or request deletion, subject to legitimate retention needs.
  • Season share links: a link works for 30 days unless you turn it off sooner. The copy behind it is deleted about 30 days after the link expires, or soon after its original expiry date if you turned it off, and at once if you delete your account.
  • Cloud AI consent: your account keeps only your latest answer, for as long as the account exists.
  • Connected assistants: an assistant's access ends as soon as you disconnect it or revoke its key. Records of that access, with tokens stored only as hashes, may be kept with your account for security and are deleted with it.
  • Waitlist applications are generally kept no longer than 24 months after your last interaction unless you become an active user or ask us to delete them sooner.
  • First-party attribution event rows are deleted after 180 days. The campaign tag stored in your browser remains until it is replaced or you clear site storage; the per-session visit marker expires with browser session storage.
  • Operational, security, and rate-limit logs are retained for limited periods, typically up to 12 months unless a longer period is needed for security, legal, or abuse-prevention reasons.
  • Subscription and transaction records may be retained as needed for accounting, fraud prevention, dispute handling, tax, and platform compliance.
  • When data is no longer needed, we delete it or irreversibly anonymize it.

10. Account and data deletion

You can request deletion of your account and associated personal data by using the account deletion option in the app settings or by contacting contact@monocursive.com.

In the app, open Settings → Account & data → Delete account, or Settings → Account & data → Delete guest account for a guest account. Share links you created then stop working, and assistants you connected lose access. In On this device there is no account on our servers: Settings → Your data → Delete everything on this device removes your data from the device. Step-by-step instructions, and what is deleted and what is kept, are on our support page: benzaiflow.com/support#delete-account.

Deleting an account removes the account and associated app data from our active systems unless we must retain limited records for legal, billing, fraud-prevention, security, or dispute-resolution reasons.

If you have an active subscription, deleting your Benzaiflow account does not automatically cancel a subscription managed by Apple App Store or Google Play. You may need to cancel it through the relevant store account settings.

If your data was shared with a processor that supports deletion or suppression requests, we will take reasonable steps to delete or de-identify the relevant data there as well.

11. International data transfers

We prioritize providers with processing in the EU/EEA whenever possible.

If data is transferred outside the EU/EEA, we rely on appropriate safeguards such as the European Commission Standard Contractual Clauses (SCCs) and supplementary measures where needed.

12. Your GDPR rights

You have rights over your personal data under GDPR. You can request:

  • Access to your data.
  • Rectification of inaccurate data.
  • Erasure of your data.
  • Restriction of processing.
  • Data portability.
  • Objection to certain processing.
  • Withdrawal of consent when processing is based on consent.

13. Complaints and supervisory authority

You can contact us first at contact@monocursive.com and we will do our best to resolve your request quickly.

You also have the right to lodge a complaint with your local supervisory authority, including CNIL in France.

14. Website analytics, attribution, and local storage

We use Plausible Analytics to understand aggregate website traffic and page performance. Plausible processes the page URL, referrer, browser, operating-system and device categories, engagement duration, scroll depth, and the destination of outbound links, and uses network information transiently to produce privacy-preserving audience counts and coarse geography. It does not place analytics cookies or provide us with a persistent cross-site visitor profile.

Our separate first-party attribution records one visit per browser session and app-store badge clicks so we can evaluate our own campaign links and aggregate acquisition funnel. An event row contains a generated event ID, event type (visit or store click), normalized campaign tag from the ref query parameter when present, landing path for visits, destination store for store clicks, timestamp, and an ISO country code supplied by Cloudflare.

The country code is coarse location derived by Cloudflare at the edge from the request IP. We do not store the source IP, user agent, account ID, cookie ID, or visitor ID in the attribution row. Attribution rows are stored in our PostgreSQL database, used only for aggregate visit, store-click, and campaign reporting, and deleted after 180 days.

The website stores the normalized campaign tag in localStorage so a later store click can be attributed, a visit-sent marker in sessionStorage to avoid duplicate events during one browser session, and your language preference in localStorage. The campaign tag remains until replaced or site storage is cleared; sessionStorage is cleared by the browser when the session ends.

The app uses on-device storage and secure storage to keep session data, preferences, drafts, notification state, and local app settings. Some information may remain on your device until you uninstall the app, clear app data, or sign out where available.

We rely on our legitimate interests for these proportionate, privacy-focused measurements. If we introduce analytics or storage that requires consent, we will request it before use where legally required.

15. Security and minors

We use reasonable technical and organizational measures to protect personal data against unauthorized access, loss, or misuse.

Benzaiflow is not intended for children under 16. If we learn that personal data of a child was provided without proper authorization, we will delete it.

16. App Store and Google Play disclosures

Apple App Privacy Details and Google Play Data Safety forms summarize selected privacy practices using each platform's required categories. Those store disclosures should be read together with this policy.

If our app features, processors, permissions, or data practices change, we will update this policy and the relevant store disclosures where required.

17. Policy updates

We may update this policy from time to time. The updated version will be published on this page with a new last updated date.