reproo

Privacy statement

Version 2026-11-13

What personal data reproo processes, why, who receives it, how long it is kept, and what your rights are.

1. Who is responsible

Wavy AS, org. no. 928 635 015, Herman Foss' gate 21, 0171 Oslo is the controller of the data about our customers and the dashboard's users, and of the operation and security of the service. Send questions about privacy to lasse@wavy.no.

When a tester reports a problem, the business that owns the app – our customer – is the controller of the data about the tester. We process it as a processor under the data processing agreement with the customer. The report form links to the customer's own privacy notice when the customer has set one, otherwise to this page.

2. Dashboard users

  • What: name, e-mail address, password (stored only as a hash), role and membership in organizations and projects, when you signed in, and the notices you chose (e-mail about new reports on or off, and the language). An invitation stores the invitee's e-mail address and who invited them.
  • Why: sign-in, access control, invitations, notices about the service and e-mail about new reports in your projects (the project name, the report's title and a link – you can turn it off under Settings). When a collaborator connects a repository to a project, the organization's owners are e-mailed who it was (name and e-mail address) and which repository.
  • Legal basis: the agreement with the customer and our legitimate interest in providing the service to the customer and the users it gives access (GDPR Article 6(1)(b) and (f)).
  • Retention: as long as the account exists. If you ask, the platform team deletes the account by hand. Accounts that only belong to a customer who has ended the relationship are deleted the same way within 30 days.

3. Customers, agreements and billing

  • What: organization number, legal name and invoice address, invoice e-mail address, reference and contact person; who accepted which version of the terms, the privacy statement, the data processing agreement and the price, and when; who ordered a project and when, whether the customer is approved, and the platform team's reason if an order was not approved; the invoice basis and whether invoices are paid.
  • Why: to enter into and perform the agreement (including checking a new customer's first order before the project is activated), invoice, follow up payment – including holds for non-payment – and document the agreement. The organization number is looked up in the Brønnøysund Register of Business Enterprises, and invoices are made in the accounting system Fiken.
  • Legal basis: the agreement (Article 6(1)(b)) and the bookkeeping obligation (Article 6(1)(c) and the Norwegian Bookkeeping Act).
  • Retention: accounting records are kept for five years after the end of the financial year (Bookkeeping Act section 13). The acceptances of the agreements are kept for as long as the customer relationship lasts and five years after, as documentation of the agreement. Orders and approvals (who ordered, when, and the reason if an order was not approved) are kept just as long, for the same reason.
  • A billing, hold, order or activation e-mail that could not be sent is kept as it was written – the recipient's e-mail address, the organization and project names and the amounts, in order and activation notices the name and e-mail address of the person who ordered or activated, and in notices of a hold or a rejected order the platform team's reason – until it is sent (new attempts for about three days), and for 30 days after we have given up sending it.

4. Testers and reported problems

We process this for the customer, as a processor:

  • What the tester writes and answers and – if the project asks for it – name and e-mail. The app can also provide a user id and a role for the tester.
  • From the app: page or screen and route, browser or device and operating system, language and time zone, app version, feature flags and the custom context the customer allows, recent actions, and summaries of network requests and console messages. Request bodies are not stored, and sensitive fields such as Authorization, cookies, passwords and API keys are removed automatically.
  • In the iOS and Android apps also the network connection (connected or not, its type, whether it is metered or constrained, and when it changed) and the time the report was written.
  • Screenshots: in the iOS and Android apps a screenshot of the screen the tester was on is attached automatically. The tester sees it and can remove it before the report is sent. On the web, screenshots are attached only when the tester does so.
  • Crash data: in the iOS and Android apps, crash data from the previous run is sent when the app starts again – the kind of crash, the error message, the call stack and the app's recent actions, screen, logs and network summaries. Crash capture is on by default and can be turned off per project; the app then deletes this data on the phone.
  • Who sees it: the customer's team in the dashboard. Code and technical details are shown only to users with technical access. The tester sees their own conversation and the report's status, never code.

The purpose and the legal basis are decided by the customer. As a tester, contact the business that owns the app to exercise your rights. If you contact us, we pass the request on and help the customer answer.

5. Storage on the phone and in the browser

  • The iOS and Android apps: a report written without a connection is kept on the phone until it is sent – for at most 7 days and at most 20 reports. It holds the text, any name and e-mail, the screenshot and the app data, is stored in the app's private area and is not included in backups of the phone.
  • Reports with follow-up questions waiting for the tester are remembered on the phone for at most 7 days (the report number, an access key and a short text from the report).
  • Crash capture keeps a small black box on the phone (recent actions, screen, logs and network summaries) that is continuously overwritten. After a crash it is sent the next time the app starts; if it cannot be sent, it waits for at most 7 days and at most 5 crash reports.
  • The apps also remember the project's settings for the report form: which fields are shown, whether crashes are captured, and the link to the privacy notice.
  • The browser: the report form remembers an unfinished conversation in the browser's local storage (the report number, an access key and the app's user id for the tester), so the tester can continue where they left off. It is used for at most 12 hours, and is deleted when the conversation ends, or the next time the report form for the same app is opened in that browser after that time. A random session id is kept in session storage and disappears when the tab is closed.

6. AI processing

The follow-up questions and the report are produced by an AI model from OpenAI (United States). Testers are told in the report form that the questions are asked by an AI assistant.

What is needed to investigate the report is sent to OpenAI: the tester's text and answers and the agent's questions, with e-mail addresses, phone numbers, national identity numbers, card numbers, account numbers (IBAN) and secrets replaced automatically before sending; technical app data without user id, name, e-mail and session id – environment, app version, page or screen and route, device, operating system and browser, language, the tester's role, feature flags and the custom context the customer allows, recent actions, summaries of network requests and console messages, and the network connection; crash data – the kind, the time, the error message, the names and messages of the exceptions in the same chain, and the app's own call-stack frames with file and line; the project's name, description, AI instructions and knowledge notes; excerpts of the customer's code; titles and summaries of similar reports and issues in the project and in the issue tracker the customer has connected, to find duplicates, and the conversation of an earlier report when the tester continues it; and the team's reason when it dismisses a report as expected behaviour, to suggest a knowledge note. Names the tester writes in the text, and what the customer puts in the custom context, may be included. Screenshots and attachments, and the name and e-mail from the contact fields, are not sent.

Requests are sent with storage turned off and are not used to train models. Under its terms, OpenAI may keep them for up to 30 days to detect abuse.

The AI makes no decisions about testers: the reports are suggestions the customer's team reviews, so the rules on automated decisions in Article 22 do not apply.

7. Code from the customer's repositories

From connected repositories, a searchable excerpt of the code (an index) is stored to find the code relevant to a report. Secrets such as API keys are removed from the index. The code may contain personal data if the customer has such data in its code. The index is deleted when the repository is disconnected.

8. Security and operations

  • To protect the service against abuse we count requests in short time windows per IP address and per session, and for sign-in, e-mail links and invitations also per account or user (the e-mail address only as a hash). The counters are deleted every night once they are more than a day old.
  • The sign-in service logs sign-ins with time and IP address. The hosting provider logs requests with IP address, time, address and status code for a short period. Our own application logs contain no tester text, AI requests or access keys.
  • When error tracking is turned on, errors in the service are sent to Sentry in the EU, scrubbed of personal data: the route, the error type and message without e-mail addresses, keys, phone and card numbers, the stack trace, the release and the environment. That includes errors in the dashboard in the browser, never the report form. IP addresses are not stored there.
  • Legal basis: our legitimate interest in keeping the service secure and stable (Article 6(1)(f)).
  • We also keep internal usage statistics (for example that a report was created or a project activated), tied to the organization and project, not to people, for 13 months.

9. Recipients and sub-processors

We use these providers, all under a data processing agreement:

  • Supabase – Database, sign-in (Supabase Auth) and file storage for screenshots and attachments. EU – Frankfurt, Germany (AWS eu-central-1).
  • Vercel – Hosting of the application: the server functions, the web pages, the report form and the scheduled jobs, delivered through Vercel's global network (CDN). The server functions run in the EU – Frankfurt (fra1). Requests pass through Vercel's edge network nearest the user, which may be outside the EEA.
  • OpenAI – The AI analysis: the follow-up questions to the tester and the report for the team (OpenAI API). United States.
  • GitHub – Access to the repositories the customer connects, through reproo's GitHub App: read access to contents and metadata, and – when GitHub Issues is the project's issue tracker – read access to issues to search for duplicates and show their status, and write access to issues, used only when a person in the team sends an issue to GitHub Issues or links a report to an issue there. United States.
  • Resend – E-mail from the service: invitations, sign-in and confirmation links, receipts and notices about invoices and holds, e-mails to the team about new reports, to owners and admins when a project stops taking reports, to owners when a collaborator connects a repository to a project (who and which repository), and alerts to the platform team. Testers get no e-mail. The e-mails are sent from the EU – Ireland (AWS eu-west-1). The API and the sending logs may be processed in the US.
  • Sentry – Error tracking, when it is turned on: errors in the servers and in the dashboard in the browser are reported to the platform team. Never in the report form, the hosted report page or the status page. EU – Germany (Sentry's EU region).

Which data each of them receives, and the basis for transfers, is set out in the data processing agreement.

Other recipients: the customer's own Jira, which the service automatically searches for duplicates with a few keywords from each report, and to which an issue is sent only when a person in the customer's team decides so (GitHub Issues is reached through reproo's GitHub App – see GitHub above); the accounting system Fiken (billing details); the Brønnøysund Register, where we look up the organization number; and public authorities when the law requires it. We never sell personal data.

10. Transfers outside the EEA

The database and the server functions are in the EU (Frankfurt), and so is error tracking (Sentry). Some providers are US companies or process data in the US: Vercel (the edge network), OpenAI (the AI analysis), GitHub (the repositories and the issues in GitHub Issues), Resend (API and sending logs), and Supabase and Sentry (operations and support). The transfers rely on the EU–US Data Privacy Framework for the providers certified under it, and on the EU standard contractual clauses (SCCs). You can ask for a copy of the transfer basis.

11. Retention

  • Reports, conversations, screenshots and attachments: until the retention period the customer has chosen for the project runs out – 30, 90 or 365 days, or unlimited; new projects have 365 days, and unlimited is the customer's own choice – and at the latest 30 days after the customer relationship has ended (the platform team deletes them by hand).
  • Tester sessions (user id and role from the app, and name and e-mail when given): until the session has no reports left and the project's retention period has run out, counted from when the tester was last active.
  • Issues the team has made from reports, and their copy in the built-in issue tracker, are not deleted with the report – they are the team's work and are kept until the customer relationship ends. The night after the report is deleted, the name of the person who reported the problem is removed from them. Issues sent to Jira or GitHub Issues stay there under the customer's own agreement with that provider.
  • The code index: until the repository is disconnected.
  • Reports on the phone: until they are sent, at most 7 days.
  • User accounts: as long as the account exists (see above).
  • Accounting records: five years after the end of the financial year.
  • Who accepted which agreements, and the orders with the platform team's approval or rejection (and the reason for a rejection): as long as the customer relationship lasts and five years after.
  • Billing, hold, order and activation e-mails that could not be sent: until they are sent (new attempts for about three days), and 30 days after we have given up sending them.
  • Operational data: usage statistics for 13 months, received GitHub notifications (webhooks) for 30 days, code indexing runs for 90 days and started GitHub connections a day after they expired.
  • Request counters (per IP address, session, account or user): at most two days – they are deleted every night once they are more than a day old.
  • Error events in error tracking (Sentry): at most 90 days.
  • Database backups are replaced continuously; what has been deleted disappears from them when the copies are replaced.

12. Your rights

You have the right to access the data about you, to have errors corrected, to have data erased, to have the processing restricted, to object to processing based on legitimate interest, and to receive data you have given us in a machine-readable format (data portability).

Send an e-mail to lasse@wavy.no. We answer within a month and may ask you to confirm who you are. The platform team handles access, hand-over and erasure by hand, following a set routine. If it concerns data we process for a customer – for example a report you sent as a tester – we pass the request to the customer and help the customer answer.

13. Complaints to the Norwegian Data Protection Authority

If you believe we process personal data in breach of the rules, you can complain to the Norwegian Data Protection Authority, Datatilsynet (datatilsynet.no, Postboks 458 Sentrum, 0105 Oslo, Norway). We appreciate it if you contact us first, so we can put right what is wrong.

14. Cookies and local storage

We use only cookies and local storage that the service needs to work, and no analytics or advertising tools or third-party cookies. That is why we do not ask for consent.

  • sb-…-auth-token: keeps you signed in to the dashboard (while you are signed in, up to 400 days).
  • reproo_device: that the browser has signed in to your account before, without your e-mail address – so failed sign-in attempts from other networks do not lock you out (180 days).
  • uatai_locale: the language you chose in the dashboard (one year).
  • uatai_last_org and uatai_project: the organization you used last and the selected project filter (180 days).
  • reproo_agreements_later_…: that you chose "Later" on the notice about new terms for an organization – one per organization (until the browser is closed).
  • The report form: local storage and session storage in the browser, as described above.

15. Changes

We update this statement when the processing changes. The version date is at the top, and material changes are notified to our customers.