Privacy notice
Version of 16 November 2026 · Zaigg Echo is in a pilot phase.
1. Who processes the data
The controller is Innovatica Technologies FZ-LLC, a company registered in the Ras Al Khaimah free zone, United Arab Emirates. For any question about your data, write to info@innovatica.ai.
2. What we process and why
| Data | Purpose | Basis |
|---|---|---|
| Account: e-mail address, role (citizen or administrator), the date you accepted the rules and terms. The sign-in code is stored only as a hash and is valid for 10 minutes. If you sign in with Google, we get only your e-mail address from Google and the confirmation that it is verified — not your name, picture or contacts. | Signing in, daily limits, preventing abuse. | Performance of the contract (Terms of use). |
| Reports and actions ("me too", "still there", fixed, reopened…): the photo, the problem type, a short note on where exactly the problem is (optional; for "Other problem" it describes the problem), the phone's GPS position and its accuracy at the time, the time, the street and place name read from the map. | A public record of problems in public space; checking the person was on the spot; a more precise problem location from several readings. | Performance of the contract; legitimate interest in an accurate public record. |
| Case e-mails: when a case you took part in changes (for example someone else says “me too”, it is marked fixed or reopened, or its photo could not be published), we send you a short e-mail in the language you use in the app. The reporter of an open case also gets one e-mail when Zaigg Echo has listed the services of its municipality, and while a case is open, its reporter and the people who said “me too” get a reminder after 3, 10 and 30 days and then monthly: to send it to the service, mark it fixed or say it is still there; one link in each reminder stops them for that case. We keep a record of what we sent so the same news is never sent twice. They are on until you switch them off (Menu → Account, or one click from any e-mail); if e-mails to your address bounce, we switch them off. E-mails about a strike or a ban (section 6) are always sent. | So you know what happened to your report and whom you can contact, and can check on the spot. | Performance of the contract; legitimate interest. |
| Places you follow: when you choose “Follow” for a municipality or settlement, we keep which place and since when, and send you one e-mail a week with what was fixed and what was newly reported there (only when there is something). You stop with one click from the e-mail or with the same button in the app; switching e-mails off stops it too. | So you can follow your neighbourhood. | Consent (you choose to follow). |
| Votes on service contacts: when you answer “Does this contact work?” for a service's contact, we store your account, the contact, your answer (works or does not work) and when you gave it. You can change your answer or take it back. | Showing which contacts work and hiding those that do not (section 5). | Legitimate interest in an accurate list of contacts. |
| Flags and takedown requests: the reason and a short explanation; for a request without an account also the e-mail address we send the receipt and the decision to. We record the decision — taken by the automatic check or, for a takedown request, by the administrator — and when it was made. | Taking down content that breaks the rules or someone's rights; a record of decisions. | Legitimate interest; compliance with legal obligations. |
| Strikes and bans: the account, the identifier of the phone the content came from, the case, whether a rule or the administrator decided, the step (a warning, or a ban and until when) and, for a ban by the administrator, the written reason (section 6). | Protecting the service and its users from misuse. | Legitimate interest; performance of the contract. |
| Messages to the team (Contact and feedback): the kind of message, its text, your e-mail address and your name if you give it; when you are signed in, your account's address instead of a typed one, the account's identifier, role and language; the page you write from. We keep them in the platform's central inbox (same controller, Innovatica), with our replies and yours. When we reply, we e-mail you from DoNotReply@zaigg.com with a private link to the conversation — anyone who has the link can read and answer it; the link lasts 90 days. | Answering your question or request; improving the app. | Legitimate interest. |
| Device identifier: a random number the app creates and keeps in your browser, sent with reports and actions. | A daily report limit per device, so one person cannot get around the rules with several accounts. | Legitimate interest (preventing abuse). |
| Technical data: IP address (only in the server's memory, for a per-minute request limit; never written to the database); server operation logs (time, the address requested without its parameters — so no positions — and errors). | Security and operation of the service. | Legitimate interest. |
Before a photo is stored we remove all its metadata, including any GPS position and device details recorded in it.
3. What is public and what is not
Public: as soon as you add a report — the problem's location on the map, the problem type, the street and place, the case's state (open, fixed or removed) and its history with dates (no times); the photo and the "Where exactly?" note once the automatic check has cleared them (section 6). If a photo was taken down, the reason is public. Who did what is shown only as “reporter” or “citizen 2”, “citizen 3” and so on, numbered within that one case, so the same person cannot be followed from case to case. For each contact of a service, the number of people who said it works and who said it does not is public.
Never public: your e-mail address, the GPS positions of individual actions, exact times, how you voted on a contact, and whether you got a strike or a ban.
Note: a report's location is where you stood when you made it, because reports are made only on the spot. The problem's public location can therefore show that someone was there that day.
4. Who else sees the data
- At Zaigg Echo no one looks at reports, photos or notes before they are published (section 6). The only person with more access than a citizen is the administrator, who deals only with takedown requests and bans and is never a routine step. For a takedown request the administrator sees the reason, the explanation, the requester's e-mail address and the photo or note concerned, and decides whether the content is removed for good or published again; the decision is sent automatically to the address in the request. The administrator also sees the list of accounts (e-mail address, date joined, number of reports, strikes and any ban) and can ban an account (for 7 or 30 days or permanently, with a written reason) or lift a ban.
- Microsoft (Azure) provides our servers, database and e-mail delivery. The data is stored in the European Union (Sweden).
- Automatic blurring and AI recognition. Right after you take the photo, a reduced copy (up to 1024 pixels, without metadata, your e-mail address or your location) is sent — before blurring — to Microsoft Azure AI Vision on our platform in Sweden (EU), which finds people, number plates and phone numbers; the app blurs them on your phone. The blurred copy then goes to one of the artificial-intelligence models we use, which looks for the other items on our blur list (for example names on doors, documents, screens, children, the inside of homes) and recognises the type of problem; the app blurs those too before the report is added. We call the models through Microsoft Azure AI Foundry; they may be models by Anthropic (Claude), OpenAI, xAI (Grok) and others, and processing may take place outside Serbia and the European Union, including in the United States, under data-processing agreements with standard contractual clauses. People are only found in order to blur them — no one is identified and no biometric templates are made. The basis is our legitimate interest in protecting the privacy of people who happen to be in a photo. The providers do not use the photos to train models; to prevent abuse they may keep them briefly, for at most 30 days. We keep the blur list and add to it when you tell us something was left visible. When the model has blurred the photo and has no doubt that it shows a public space with no face or number plate still visible, the photo is public at once; if it is unsure, or the automatic check fails, the photo stays hidden and the case is public without it — no person looks at it (section 6). When a signed-in user flags a photo, the stored (already blurred) photo is sent once more, without your e-mail address or location, to a different one of the models we use for a second look; as with the first check, the copy is not kept — only which model answered and what it found. Automatic blurring can miss something — you can flag a photo or ask for it to be taken down. A photo taken on the spot to mark a case fixed or to reopen it goes through the first step only (Azure AI Vision in Sweden): if it works, the photo is public on the case; if not, only the people on the case and the administrator see it. The short note with a report ("Where exactly?") — the text only, without the photo, your e-mail address or location — is checked by the same models before it is published, so that it carries no details about people, insults, politics or adverts.
- Finding public services (the census). To list a municipality's public services (section 5), an AI model with web search — Claude (Anthropic), through the same Microsoft Foundry service — looks for them on public web pages. It is given only the municipality's name and country and the kinds of problems, never anything about you or your report. Our server then opens each page itself and keeps a contact only if it really appears there.
- When you open the app, your browser itself downloads the map from OpenFreeMap, fonts from Google Fonts and the map library from cdnjs (Cloudflare). These services see your IP address, and OpenFreeMap also sees the part of the map you view.
- Searching the map. When you look for a place or a street, our server sends the words you typed and the part of the map you are looking at to the OpenStreetMap Foundation's place search (Nominatim) — without your IP address, account or location. A case number is looked up only in our own database. We do not keep searches with your account.
- Signing in with Google (if you choose it instead of a code): you sign in at Google, which sends us only your e-mail address and the confirmation that it is verified. Google then knows that you signed in to our app, under its own privacy policy. Signing in with a code by e-mail is always possible.
- When you sign in, when you send a takedown request and when you send the team a message without signing in, Cloudflare Turnstile checks that you are not an automated program; Cloudflare sees your IP address and basic information about your browser. The basis is our legitimate interest in protecting the service from fake accounts and misuse of the form.
We do not sell or rent personal data. There are no ads. The app has no tracking cookies.
5. Contacts of public services
Zaigg Echo does not contact services: it sends them no cases or messages of any kind and does not read their replies. Instead, for each municipality where people report problems, it keeps a public list of the public services there (the municipality's offices, public utilities, inspections, road authorities and the like) with their public contact details, so that people can contact them themselves.
- What is listed: each service's e-mail addresses, phone and mobile numbers, WhatsApp and Viber numbers, website, online forms and street address, as the census (section 4) found them on official web pages: the service's own website, the municipality's website or a public register. Each contact is shown with the page it comes from and the date it was last checked, and only if Zaigg Echo's own visit to that page found it there. The list is published automatically, without anyone approving it.
- Contacts of named people: a contact tied to a named person (for example an inspector's mobile number or a clerk's e-mail address) is listed only when the service's own official page gives it as a contact for the public — never from news, social networks or other sources. The basis is the public's legitimate interest in knowing how to reach the public services that look after public space, using details the services themselves published for that purpose. If a listed contact names you, you can object and ask for it to be removed by writing to info@innovatica.ai; we take it off the list.
- “Does this contact work?” Signed-in people can say whether a contact works (a wrong number, an e-mail address that bounces or a wrong office does not). We store who answered what (section 2); the numbers are public, who answered is not. When at least 3 different people who have taken part in Zaigg Echo (with an accepted report or a check on the spot) say a contact does not work, and they outnumber those who say it does, the contact is hidden and the municipality's services are looked up again (at most once in 30 days).
- When you contact a service, you do it yourself: tapping a contact opens your own e-mail, phone, WhatsApp or Viber, with the case's text and link if you started from a case. The message goes from you, through your own provider and under its terms; Zaigg Echo does not see it, and the service sees who you are (your address or number).
6. Automatic decisions
No person reviews reports before they are published. The following is decided automatically, by the checks described in section 4 and by fixed rules:
- Photos: a report's photo is published when the automatic check clears it. If the check is unsure, the photo stays hidden and the case is public without it; the stored photo is checked again for up to two days, and if it is still not cleared it stays hidden and we tell you once that a new photo helps. No person reviews a hidden photo: the way back is a new photo taken on the spot. A photo the check finds unusable (for example a selfie or a person as the subject) cannot be used.
- Notes: the "Where exactly?" note is published when the check accepts it; a note it rejects cannot be added, and one it is unsure about is put to other models and stays hidden unless they all accept it.
- Flags: a flag hides the photo or note at once. A different model then looks at the flagged photo or note again: if it confirms the problem, it stays down; if it finds it clean, it comes back; without an answer, it stays down. A takedown request also hides the content at once and then waits for the administrator.
- Strikes and bans: an account gets a strike when the automatic check rejects something it adds as misuse: a photo with offensive content, advertising or spam, or a note with an insult or an advert. A political or unrelated note is only refused, without a strike, and nothing gives a strike merely because a photo is blurred, unclear or unusable. A flagged photo gives a strike only when the second look itself finds such misuse — never for a mismatched problem type, visible faces or number plates, or for showing someone's property. Flags alone, however many, never give a strike. An account gets at most one strike per case and at most one a day. Each strike within 365 days — counted together with those on the phone the content came from — moves the account one step: a warning, then a 7-day ban, then a 30-day ban, then a permanent ban. A ban covers that phone too. While banned you can still look at the map and cases, answer on your own cases (still there, fixed or not, withdraw) and leave any case, but you cannot report or take part in other people's cases.
These are decisions based solely on automated processing that affect you (Article 22 of the General Data Protection Regulation, GDPR, and the corresponding rules of the law that applies to you); they are necessary to provide the service under the Terms of use. We tell you about every strike and ban by e-mail: what happened, what it means and what the next step would be. You can ask for a person to review a strike or a ban, give your point of view and contest it: write to info@innovatica.ai, with the report number if you have it. The administrator can lift a ban or a strike; a lifted step no longer counts.
7. How long we keep data
- Your account until you delete it. A sign-in session lasts 30 days.
- Reports and actions are kept as the public record while the case exists, including each action's GPS position — it only makes the problem's location more precise and is never published. When you delete your account they stay, with no link to you: we delete the link to the account and the device identifier.
- The device identifier stored with an action (for the per-device daily limit) is deleted after 90 days.
- Photos that are not public — taken down after a flag or a takedown request, belonging to a removed report, or never cleared by the automatic check — are deleted 30 days after that decision, and a withdrawn report's photo 30 days after the withdrawal; a photo hidden only by a flag or a takedown request that has not been decided yet is kept until it is decided. The case and the reason stay on record. Removed reports are not public.
- Flags and takedown requests are kept as the record of decisions for as long as the case exists.
- Blurring and AI recognition: we do not keep the copies sent to the detector and the model — only a fingerprint (SHA-256, so the same photo is not paid for twice), which service or model answered, where it found what to blur (rectangles, no image), what it suggested, how long it took and what it cost; with a report, also which suggestion you saw, to measure the models' accuracy.
- Strikes, warnings and bans: the record described in section 2. The phone identifier stays in it beyond 90 days and if the account is deleted, so that a ban cannot be got around with a new account. A strike counts toward the next step for 365 days; the record is deleted two years after the ban ends (a warning ends the day it is given), and the record of a permanent ban is kept while the ban lasts. The basis is our legitimate interest in protecting the service from misuse.
- Votes on service contacts are kept while the contact exists; they are deleted when you take your vote back or delete your account.
- Contacts of public services are kept while the service is listed; a municipality's census is redone after a year where people still report, and sooner when people say a contact does not work. A contact that names you is removed when you ask (section 5).
- Records of features that no longer exist (the e-mails Zaigg Echo sent to services and the services' replies, the links in the digests sent to services, people's suggestions and corrections of which service handles a case, people's standing and reputation in each municipality, the log of role changes and requests for service accounts), were deleted when those features were removed (November 2026) and are no longer kept.
- Messages to the team are deleted 24 months after the conversation ends (or after its last change, if it never ends), with the replies; messages rejected as spam within a day.
- Server logs for at most 30 days.
8. Cookies and data in your browser
We use only necessary cookies: civi_session, which keeps you signed in (30 days), and when you sign in with Google the short-lived civi_session_google (10 minutes), which ties the return from Google to your browser. In your browser we keep your chosen language, theme, map colouring and the device identifier.
9. Your rights
You have the right to access your data, to rectification, erasure, restriction of processing and portability, and to object to processing based on legitimate interest. You can delete your account yourself in the app (Menu → Account → Delete account). If you recognise yourself or your property in a photo, send a takedown request (no account needed): we take the photo down at once and e-mail you the decision. For any other request write to info@innovatica.ai.
You can also object to a contact that names you being listed (section 5), and you have the right not to be subject to an automatic strike or ban without being able to ask for a person to review it, give your point of view and contest it (section 6).
You have the right to complain to the Commissioner for Information of Public Importance and Personal Data Protection (Serbia), Bulevar kralja Aleksandra 15, 11000 Belgrade, www.poverenik.rs.
10. Age
Zaigg Echo is for people aged 16 and over.
11. Changes
When this notice changes materially we will update the version date and ask you to accept it again before your next report.