Privacy policy
How heyrup.app handles your data
Last updated: 15 September 2026
1. Service and contact
heyrup.app is operated by Radek Duha, Czech business ID 01091484, with a registered address at Tulipánová 2805/3, Záběhlice, 106 00 Praha 10. The application helps business owners find PDF attachments in email, read their contents, recognise documents and assign them to businesses according to confirmed rules or individual user decisions. It presents uncertain results for review. The aim is to make preparing documents for accountants easier.
Send questions and requests for data access or deletion to jsem@radekduha.cz. This policy applies to both the public website heyrup.app and the application app.heyrup.app.
2. How we access your mailbox
For IMAP, enter the server address, username and password, or an app password created with your provider. The connection uses encrypted TLS transport on port 993. We store the password encrypted with AES-256-GCM; the service uses it to access the mailbox. Once the connection is verified, choose the folders HeyRup should search.
HeyRup opens selected folders in read-only mode and downloads PDF attachments within the selected period. It does not move, delete or mark messages as read. The credentials may allow broader access at your provider; read-only mode limits our processing, not the permissions granted by the password itself.
3. What data we process and why
When you start a search, we retrieve matching messages and their structure. For PDFs found, we record the subject, sender, date and filename so you can trace their source. During processing, we download PDFs and recognise their content. IMAP reads metadata and selected PDF message parts; we do not store email bodies in the document database or send them for AI classification.
We retain the connected mailbox’s email address, message and attachment identifiers, subject, sender, date, filename, file type and size, a copy of the PDF, extracted OCR text and the recognition result, including its reasoning and confidence level. A document may contain names, addresses, business and tax identification numbers, bank details, amounts and purchase information. We also store attachments marked as non-invoices or unclear so that the search results remain traceable.
For repeated IMAP access, we store the server address, username, encrypted password and selected folders. For any Google OAuth connection, we store an encrypted refresh token. We also process the app account ID and email, organisation and businesses, rules, document decisions, recipient contacts and delivery settings. For deliveries, we record files sent, attempts and outcomes. We keep technical request and processing logs for troubleshooting.
4. OCR, AI and technical providers
Data is processed on servers, not only in your browser. Railway hosts the application and database; we store PDF copies in Google Cloud Storage. When recognising a document, we send the PDF to Google Document AI to convert it to text (OCR). This involves more than reading the attachment filename or email subject.
We send the extracted document text to the OpenAI API to identify the document type and the information needed to assign it to the correct business. If you describe a new rule in your own words, OpenAI receives that description and the details of the business you selected to draft a rule. The draft takes effect only after your confirmation. This is not training a general-purpose model; automatic recognition may make mistakes, and uncertain results require review.
Clerk handles application sign-in. Cloudflare serves the public website. Signing in to the app and connecting a mailbox are separate steps. Paid plans are in preparation; the following description also covers data flows once they are available.
We use Resend for document delivery. When sending, this service receives your chosen address, message and the relevant PDFs. Manual delivery requires confirmation of a specific preview; scheduled delivery follows the schedule you have explicitly enabled. Saving a recipient address alone does not send any documents. HeyRup sends messages through its own provider, not through your mailbox.
Stripe is prepared for subscription and payment management. When you use a paid plan, it processes the information needed for payment and billing, not the contents of email attachments or document OCR text. You enter your payment card details directly in Stripe.
5. Data use and the planned Gmail connection
We use mailbox data only for the described heyrup.app features. We do not sell it or use it for advertising, ad profiling or training general-purpose AI or ML models. We limit disclosure to technical providers to delivering these features, securing the service and meeting legal obligations.
People must not browse your data except with your explicit consent for specific support, for a necessary security investigation or to meet a legal obligation. The Gmail connector is not publicly available yet. Once released, connection will be optional through Google OAuth with openid, email, profile and gmail.readonly permissions; the latter technically allows reading the whole mailbox. This connection does not give us your Google password. The following Limited Use commitment also applies to data obtained through Google APIs.
heyrup.app’s use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements.
6. Retention, disconnection and deletion
We retain working copies of PDFs, OCR text and content recognised from documents for three calendar months from the first import. After this period, they are no longer available to open or deliver, and we remove them from the active environment. Reprocessing or sending does not extend the period. HeyRup is not a long-term accounting archive; arrange the required document retention separately.
You can delete a working copy earlier directly from the document in the app. Stored objects are removed by follow-up jobs, which retry if a technical error occurs. We do not delete the original email or attachment in your mailbox. Deleting a copy in HeyRup also does not remove files you have already downloaded or delivered to another recipient.
We retain delivery history, including recipient addresses, attachment filenames, sending attempts and provider events, along with activity records, for six months and then delete them. File fingerprints and source identifiers remain for the lifetime of the organisation to prevent expired documents from being unintentionally restored, charged again or delivered twice. We retain identifiers and counts of consumed plan units for the same period. These remaining records do not contain PDFs, OCR text, or line items and amounts from customer invoices.
Disconnect an IMAP mailbox in HeyRup to remove its stored credentials and stop further access. For any Google OAuth connection, you can also revoke consent in your Google account connections. Revoking consent stops further authorised access, but does not itself erase documents already stored, your account or app settings. It also does not cancel a paid plan or delivery of documents already stored; you manage those separately in the app. To remove connections, tokens, your entire account and associated data, contact jsem@radekduha.cz.
PDF storage in Google Cloud is private and located in the EU. After an object is deleted, it has a seven-day protection period for technical recovery; during this period, the document cannot be opened or delivered in HeyRup. This location does not automatically apply to all the service’s other providers.
We store our database backups in separate, private Google Cloud storage in the EU. They are configured for deletion after seven days, followed by a further seven days of technical recoverability after deletion. Removal is asynchronous, so this is not a guarantee of erasure on exactly the fourteenth day. Deletion from the active application does not mean immediate erasure from backups or providers’ operational logs. We will address the scope and completion of a requested erasure with you through the contact above.
7. Data protection
We control document access by signed-in user and organisation. An accounting firm's recommendation does not give it access to a client's mailbox. We store IMAP passwords and any Gmail refresh tokens encrypted. The website uses HTTPS and IMAP connections use verified TLS. Providers receive the data needed for their specific processing; no security method eliminates every risk.
8. Your requests and rights
Connecting a mailbox is optional and you can disconnect it in HeyRup. To request access to, correction, export, erasure or restriction of your personal data, contact us at jsem@radekduha.cz. Before handling a request, we may reasonably verify that it concerns your account. This does not limit your rights under applicable regulations.
9. Policy changes
We publish the current wording and update date at this address. Before introducing a new processing purpose or different data sharing, we will update the information for users and request consent where required.
HeyRup launch notifications
If you ask for a launch notification, Radek Duha stores your email, the form language and the date of your consent, based on that consent. We use this information only to announce the launch, not to subscribe you to a newsletter. It is stored in a private Cloudflare database; Cloudflare Turnstile protects the form.
Every day, we remove records older than 180 days from the active database. The provider’s technical backups may retain data for longer.
You can withdraw consent and request deletion at any time by emailing jsem@radekduha.cz.