Every camera system answers one question before it answers any other: who is this for?
A system built for a property owner watches that owner's property, stores video the owner controls, and exists to answer questions the owner has a legitimate reason to ask. A system built as infrastructure watches everyone who passes, stores what it captures in a database the property owner has never seen, and answers questions asked by people the property owner has never met.
Both are made of cameras. They are not the same thing, and the difference is not the hardware — it is who holds the keys.
Duluth Security Needs builds the first kind. This page is the written version of that commitment, so you can hold us to it and so you can measure any other vendor against it.
What we commit to
1. You own the footage
Video recorded by your system is your property. It records to a server on your premises, on hardware you paid for and hold title to. We do not keep a copy.
2. You own the hardware, the software configuration, and the credentials
At the end of every project you receive documentation: device inventory, IP addressing, every username and password, and the configuration required to maintain the system. Nothing is withheld to keep you dependent on us.
3. No vendor cloud is required for your system to work
Your cameras do not need our servers, an outside company's servers, or an active subscription to record and retain video. If we went out of business tomorrow, your system would keep running. Off-site backup and remote viewing are available when you want them, configured as options you elect — never as a condition of the system functioning.
4. Your system is not connected to any shared network
There is no regional database, no cross-agency search tool, and no "network effect" behind a system we build. Your cameras are searchable by your people and no one else. We do not participate in, and will not connect clients to, systems that pool video or vehicle data across unrelated organizations.
5. Law enforcement requests go to you, not to us
We hold no footage, so there is nothing for us to hand over. If a request, subpoena, or warrant arrives, it arrives at your door, and the decision about how to respond is yours to make with your own counsel. We think that is where the decision belongs.
6. We do not sell or operate license plate recognition on public roads
We do not install automated license plate recognition aimed at public streets, and we do not connect client systems to ALPR networks. Reading the plate of every car that drives past your building is a different business from securing your building, and it is not one we are in.
7. Remote access is yours to grant and yours to revoke
If you contract us for ongoing support, that access runs through a named account you can see, audit, and disable without calling us first. We do not retain standing access to systems we no longer support, and we do not build in undocumented accounts. We have been hired to clean up after consultants who did both, and we will not become the next one.
8. Detection runs on your hardware
The AI that distinguishes a person from a passing car runs on the server in your building. Video is analyzed where it is recorded. It is not uploaded to an outside company for processing, which means there is no third party accumulating footage of your property, your staff, or your customers as a byproduct of your alarm working correctly.
9. We design to the narrowest scope that solves the problem
Cameras get aimed at what the problem is, not at everything within reach. A loading dock problem gets a loading dock camera. If a camera would cover a break room, a neighbor's window, or a residential unit's interior with no security reason to do so, we will say so during the site survey and propose a different placement.
10. Retention is a decision, not a default
We set retention deliberately, based on how long you actually need to be able to answer a question — usually somewhere between two weeks and ninety days. Longer is not automatically better. Video you no longer need is a liability you are still storing.
What we will not build
We turn down work. Specifically:
- Systems whose evident purpose is to monitor a specific individual rather than to secure a property.
- Camera placements intended to surveil the interior of a residence a client does not live in, beyond legitimate common-area coverage.
- Configurations designed to conceal from employees that they are being recorded, where notice is required.
- Any deployment where the described purpose would put us in the position of helping someone invade another person's privacy.
We would rather lose a sale than install something we would not want pointed at us.
Capability, and the limits we put on it
Modern detection is genuinely powerful, and pretending otherwise would be dishonest. Systems we build can distinguish people from vehicles from animals, ignore the blowing snow and headlight sweep that make conventional motion alerts useless, and be trained to recognize objects and conditions specific to your site. We write custom software when an off-the-shelf product cannot do what a client needs.
That capability is exactly why the scope rules above matter. A few examples of how we apply them:
Escalating response, not constant watching. Rather than a person staring at monitors, a system can respond in stages — a light, then an announcement, then a siren, then a notification to a specific phone — with each stage triggered by a defined condition. Nobody is watching your parking lot at 2 a.m. The system is, and it only involves a human when a rung on the ladder is reached.
Escort mode. A staff member walking to a car after close can trigger a monitored window: the system tracks the walk, and if the expected sequence does not complete, it escalates. It exists for the length of the walk and then stops.
Biometrics stay local, and stay optional. Where a client has a legitimate need for facial recognition — a controlled-access area, a barred-individual list — the template database lives on the client's server, is never transmitted off-site, and is never shared with us or anyone else. We require a written policy and appropriate employee notice before deploying it, and we do not lead with it, because most access control problems are better solved with a credential than with a face.
Live monitoring is scoped and scheduled. Our live video monitoring service covers defined hours and defined conditions — typically overnight human detection on a property that should be empty. It is not open-ended observation of your business during business hours.
Questions worth asking any vendor
Whether or not you hire us, these questions will tell you what kind of system you are being sold. Ask them before you sign, and ask for the answers in writing.
- Where is my video physically stored, and who else has a copy?
- If I stop paying you, does my system keep recording?
- Is my footage searchable by anyone outside my organization? By any government agency?
- Who processes the AI detection — my hardware, or your servers?
- Will I receive every password and configuration file at the end of the project?
- What happens to your access to my system when our contract ends?
- If someone subpoenas my footage, do they come to me or to you?
- What is the default retention period, and can I change it?
A vendor who cannot answer these plainly is telling you something.
Why we publish this
Security work runs on trust, and trust that is only spoken is trust you cannot check. Putting our commitments in writing means a property manager can hand this page to a tenant, a parish council can hand it to a congregation, and a business owner can hand it to an employee who asked a fair question about the new camera in the hallway.
It also means that if we ever fall short of what is written here, you have the receipt.
If you want to talk through how these commitments would apply to your building, we do free site surveys across the Twin Ports. See our commercial security camera systems, residential systems, or get in touch — (218) 228-7778.