Security and data: questions from city IT staff
Last updated November 12, 2026
Short answers for the people who look after a city's computers and records. They describe how Munitown works today. Munitown is provided by GlessTech LLC, a Michigan limited liability company doing business as Munitown.
We say what we do and what we do not do, so you can check each answer. To send us a security questionnaire, email security@munitown.com.
Where does it run?
- Munitown runs on Cloudflare, as one Cloudflare Worker: a single program that serves every city's resident site and the admin area.
- Reports, settings and users are kept in a Cloudflare D1 database. Photos are kept in Cloudflare R2 storage.
- Cloudflare encrypts the database and the photos where they are stored (AES-256), and they travel over encrypted (TLS) connections. Cloudflare manages the encryption keys.
- Browsers are told to use only encrypted (HTTPS) connections to every Munitown site for a year (HSTS).
- Cloudflare chooses where on its network the database and photos are stored. We have not limited storage to one country or region, so we do not promise one.
- Cloudflare also runs the check that keeps bots from sending reports (Turnstile), and sends our emails (Email Service).
- Cloudflare keeps a log of each request, such as the page asked for, the visitor's internet address and the response, for 7 days (Workers Logs). We use it only to find and fix problems.
- When something fails, the same log holds the error. It leaves out the subject of an email that could not be sent, since a subject can name a problem and its location.
- If a resident presses "Use my current location", we send Geoapify that position, rounded to about 10 meters, to find the street address. For a city that has set its map area, we also send Geoapify what was typed as the location of a report with no position in or near the city, when it starts with a house number, whether a resident or staff typed it, to place it on the map, the address on a report when staff press Find on map, and the city and state typed in Settings. The admin area's map loads its map images from OpenStreetMap, which sees the staff member's internet address.
- The test and demo sites have their own databases and photo storage, apart from the one cities use.
Which other companies handle a city's data?
- Cloudflare: everything above. It offers a data processing addendum and lists the companies it uses.
- Stripe: payments. We send it the city's name, the number we give the city, its billing email and purchase order number. It never gets anything from a resident's report, and we never see card or bank account numbers.
- Geoapify: only the rounded position and the typed location described above, never who sent a report or its description.
- OpenStreetMap: only the map images in the admin area.
- Google: our own email. It holds trial requests, feedback from city users, the alerts we are sent when something fails, and what a city emails us. We do not send it residents' reports.
- No one else. The pages load no analytics, ads, fonts or scripts from other companies, apart from Cloudflare's bot check on the report form and the trial form.
Who can see a city's reports?
- Signed-in city users, whom the city adds on its Users page, and our own platform administrators, who look after every city.
- Each user's access is checked against the database on every request, so removing a user or changing their role takes effect at once.
- An Administrator can do everything: departments, problem types, settings, billing, users and the data export.
- Staff work the reports: the inbox, each report and its photos, and updates. Staff given a department see only that department's reports, plus any report assigned to them.
- Residents' photos are shown only to signed-in city users.
- Anyone with a tracking code sees only the report's status, its dates, the kind of problem, messages the city chooses to share, and a photo of the finished work if staff choose to share one. Never the description, location, resident's photo or contact details.
- A department can be given an email address, which may be anyone's, such as a shared mailbox. It is emailed each report that arrives for that department or is moved to it: the kind of problem, where it is and what the resident wrote. Those emails leave out the resident's name, email and phone number.
How are cities kept apart?
- Every city's reports are in one shared database, not a database each. Each report belongs to one city, and the program loads a report only through the city it belongs to.
- A user who opens another city's admin pages is turned away, and is told nothing about whether that city exists.
- Automated tests try every city admin page as a user from a different city, and check that Staff cannot open pages meant for Administrators. They run on every proposed change to the program.
How do city users sign in?
- There are no passwords. A user enters their email, and we email them a 6-digit code and a link. Each one works once, and stops working after 15 minutes.
- Signing in keeps them signed in for 14 days, with a cookie that the page's scripts cannot read and that is sent only over encrypted connections. The session is kept on our side, and signing out ends it.
- Our database stores only a one-way fingerprint (a SHA-256 hash) of each code, link and session, never the originals.
- Wrong guesses at a code, sign-in attempts and sign-in emails are all limited.
- An email address that is not a user gets the same page as one that is, and no email.
- There is no single sign-on (SSO) and no second sign-in step beyond the emailed code.
How is the site protected from attacks?
- What people type is checked before it is used, passed to the database as values and never as part of a query, and escaped when it is shown on a page.
- A content security policy lets pages run only our own scripts and Cloudflare's bot check, and stops other sites from showing our pages in a frame.
- The admin area, the sign-in forms and a resident's report form accept a form only when it was sent from our own pages (a cross-site request forgery check).
- A report must pass Cloudflare's bot check, which our server verifies, and we limit how many reports one internet address can send.
- A photo can be up to 10 MB. The server works out its type from the file itself, not from what the browser says, and turns down a file it cannot read. Residents' photos are served only to signed-in city users.
- The export guards its spreadsheets against cells that Excel would run as formulas.
- The keys for the services we use are kept as Cloudflare secrets, not in the program's code. Messages from Stripe are accepted only with a valid signature.
What is recorded about who did what?
- Each report's timeline records every status change, assignment, note, message to the resident, removal of a reporter's details and deleted photo, with the time and, when a person did it, the user. The export includes it (history.csv).
- We are emailed when a page fails with an error, an email cannot be sent, or an hourly job fails, such as the one that removes reporters' details.
What do you keep about residents, and for how long?
- What the resident sends: the kind of problem, where it is, what they wrote, a photo if they add one, and their name, email and phone if they give them.
- The phone or computer usually makes a new copy of a photo before sending it, which leaves out its location data. If it cannot, the photo is sent as it is. Either way, our server removes the location and camera details (EXIF and XMP) from every photo, in every format. It turns down a photo it cannot read.
- We use a resident's internet (IP) address to stop repeated or automated reports, and do not store it with the report. It is in the request log above for 7 days.
- By default everything is kept, because records schedules differ. Under Settings, an Administrator can have reporters' names, contact details and photos removed 30, 90, 180 or 365 days after a report is closed, or remove them from one report at any time.
- Removal takes away the name, email, phone and the resident's photo. The location, the description and the timeline stay, so a description that names a person still does.
How are backups done?
- The database uses Cloudflare D1's Time Travel, which is always on. On our Cloudflare plan (Workers Paid), it can put the database back the way it was at any minute in the last 30 days.
- A restore puts back the whole database, for every city at once, and replaces what is in it now.
- Time Travel covers the database, not the photos.
How does a city get its data out?
- An Administrator opens Settings, then Export reports, and downloads one zip file.
- It holds every report (reports.csv), every timeline entry (history.csv), the photos, and a README.txt that explains each file. The spreadsheets open in Excel.
- A city with many photos gets them in a few numbered downloads.
- Only the city's Administrators and our platform administrators can download it. It still works after a city's account is archived.
- The export is free. To have a city's data removed for good, see "A city's data" in our terms of use.
What does an outage look like?
- The resident site and the admin area are one program, so when it is down, both are: residents cannot send or check reports, and staff cannot work them.
- Keep your city's phone number and emergency notice on your own website too, so residents can still reach you.
- If a resident sends the same report form twice, for example on a slow connection, it makes one report, not two.
- We cannot promise the site will always be available. See our terms of use.
What if someone gets into a city's data?
If we learn that someone got into a city's data without permission, we tell its Administrators promptly, say what we know, and keep them updated. Our terms of use promise this to every city that subscribes.
How do I report a security problem?
Email us at security@munitown.com. Security tools can find the same address in our security.txt file.
If you look for a security problem in good faith, do not look at or change data that is not yours, and give us a fair chance to fix it before you tell anyone else, we will not take legal action against you for it.