Frigate alternative without Docker for Windows, Ubuntu and Raspberry Pi

You have a Docker Compose file open, a YAML configuration file beside it and perhaps a Coral detector on order, but what you need is simple: get the security cameras recording. Maybe the available machine is a Windows PC rather than a Linux host. Maybe you would rather use a Raspberry Pi without making Home Assistant the centre of the system. If you are searching for a Frigate alternative, our answer is NetMagen Vision. You install it as a normal application, scan the network and start recording to your own disk.

Why seek a Frigate alternative?

Frigate deserves respect. It is a free, open-source NVR (network video recorder) built around real-time AI object detection, with an active community and close Home Assistant integration. For a self-hosting enthusiast who enjoys building around Docker, it offers a capable and flexible foundation.

That foundation also defines the work involved. The recommended installation uses Docker with Docker Compose. Configuration lives in YAML, although the web interface provides an editor for that file. Detection is designed around a selected detector such as Coral, OpenVINO, another GPU option or Hailo on Raspberry Pi. Choosing and tuning that path is part of the system build.

The difference is not that one approach is serious and the other is not. They are designed for different desks. Frigate suits an owner who wants the recorder to become part of a wider Home Assistant or MQTT environment. We built NetMagen Vision for the person who wants a native recorder working on the computer already available, without a container stack, a configuration file or an accelerator purchase.

The mobile path differs as well. Frigate has no official native mobile app. It can run as a progressive web app, unofficial apps are available, and the owner arranges remote access through a VPN, reverse proxy or Home Assistant remote access. NetMagen Vision provides a native iPhone app and, with a paid NetMagen account, a managed remote relay when remote viewing is enabled.

NetMagen Vision in one screen

NetMagen Vision turns an existing computer into a continuous recorder for existing IP cameras. The working sequence is deliberately short:

1. Install. Use the native package for Windows, Ubuntu or supported Raspberry Pi hardware.
2. Scan. Open Scan Network. NetMagen Vision discovers ONVIF cameras on the local network and tests each stream.
3. Record. Enter the camera password, add the working stream with one click and choose the drive that will hold the recordings.

The result is visible immediately. Camera tiles appear. The timeline starts to fill. NetMagen Vision records continuously to the selected local disk and starts with the computer from then on. There is no account requirement for local use and no camera limit imposed by the application.

ONVIF handles discovery, while RTSP provides broad camera compatibility: NetMagen Vision works with any camera that offers an RTSP stream. Cameras, zones and settings are managed in the application's own screens rather than in YAML.

Can I record cameras without Home Assistant?

Yes. NetMagen Vision is a standalone desktop NVR and does not require Home Assistant. Open the live wall to see every camera in a grid, or drag a mosaic video wall onto another monitor. Full-screen viewing, labels and audio are controlled from the application. The recorder remains useful as a recorder, without asking you to build an automation platform around it.

Frigate on Windows versus native installation

Running Frigate on Windows usually means adding another layer. Windows is not officially supported in Frigate's installation documentation. Some users run it through WSL or VirtualBox, but Frigate warns that passing a GPU or Coral through those environments may be difficult or impossible.

NetMagen Vision takes the native route on Windows 10 or 11, 64-bit. Download the normal installer and run it without administrator rights. There are no drivers to install, no Docker environment to prepare and no account to create. GPU acceleration is optional and can be installed later from Settings; local detection does not depend on it.

Once cameras have been added, NetMagen Vision starts with the computer. Recording resumes as part of the normal Windows startup without requiring a container host, virtual machine or hardware passthrough. The practical gain is easy to picture: restart the PC, return to the application and see the day's timeline filling again.

Linux NVR software without Docker

NetMagen Vision is Linux NVR software delivered as a native .deb package for Ubuntu 22.04 or newer on 64-bit x86. One apt command installs it. There is no container image to manage and no YAML configuration file to write.

The Linux package opens to the same pages and settings used on Windows and Raspberry Pi. Scan Network discovers cameras. The live wall shows them. Zones are drawn on the camera view, and storage is selected inside the application. Moving between platforms does not mean learning a different recorder interface.

GPU acceleration comes pre-installed in the Linux x86-64 package. It is available without making a dedicated accelerator stick a prerequisite, while recordings and local object detection continue to stay on the machine.

Can a Raspberry Pi NVR skip accelerators?

A NetMagen Vision Raspberry Pi NVR can perform local object detection on the CPU. No Coral is required. The supported path is a Raspberry Pi 5 or Raspberry Pi 4 with 8 GB, running 64-bit Ubuntu 24.04 or newer and using the native .deb package.

Recording itself is light because NetMagen Vision copies the camera streams rather than decoding them. The Raspberry Pi can therefore record as many cameras as its network allows. Local live viewing is a different workload, so the interface displays a few tiles at a time on the Raspberry Pi itself.

Frigate can also run on Raspberry Pi, but detector setup needs additional care. Its documented considerations include limited GPU memory, a separately powered USB hub for a USB Coral and operating-system caveats for Hailo drivers. That work is reasonable when it is part of the project you want to build.

With NetMagen Vision, the Raspberry Pi remains a straightforward recorder: install the package, discover the cameras and let CPU detection distinguish people from vehicles and animals. There is no accelerator stick or powered hub to add before the first detection marker appears on the timeline.

The Frigate mobile app question

The Frigate mobile app question often turns into a remote-access design question. Frigate has no official native mobile app, although it can be installed as a progressive web app and unofficial third-party apps exist. Access from outside the home is arranged by the owner through a VPN, reverse proxy or Home Assistant remote access.

NetMagen Vision provides a native iPhone experience through the NetMagen Vision iPhone app, as well as browser access at netmagen.com. With a paid NetMagen account and remote access enabled, you can open the live view, select a date, inspect the same full-day timeline and cut a range to MP4 without re-encoding. You can be elsewhere in the city, open last night's archive and jump directly to the relevant motion, sound or transcript entry.

The recorder makes an outbound connection to the NetMagen relay. No port forwarding, open home ports or router configuration is required. The connection is certificate-pinned, every request is signed, and the relay forwards the encrypted stream without keeping a recording copy. HLS is the default, with WebRTC available when lower latency is wanted.

The iPhone app also has a free Local mode for viewing cameras directly on the same Wi-Fi without an account. It contains no ads, analytics or tracking. Mobile is iPhone only today, but alerts are sent separately through Telegram and arrive on any device that runs Telegram, including Android phones. The app does not generate push notifications; the recorder sends the Telegram messages.

Free tools and practical comparison

NetMagen Vision gives you a timeline designed for finding an event rather than merely proving that a file exists. A full day sits on one strip. Motion is blue, sound is red, and speech and detections appear as markers. Scrub across the strip, jump to a moment, play at up to 16x and export the useful clip. Speech is transcribed, and selecting a transcript line seeks directly to that point.

Alerts are equally concrete. Draw a zone on a camera. When someone crosses it, Telegram receives a message with a snapshot. Local object detection distinguishes people, vehicles and animals on the recorder itself, without sending that detection work to a cloud service and without requiring an accelerator stick.

The recording, timeline, transcripts, local detection and Telegram alerts are free, with no account and no camera limit. NetMagen Vision is freemium rather than open source, and it lists the open-source components it uses. Remote viewing, Vision AI, shared access, the managed relay and priority support require a NetMagen account because they use NetMagen services. The NetMagen Vision product page explains the free and paid split.

Vision AI can recognise faces and licence plates by name, maintain a gallery and watch list, present a review queue and open the recording at any sighting. Shared access lets the owner invite family or staff, group them and control access to cameras and features, including live video, recordings, exports, transcripts, Vision AI, schedules and expiry. The owner can also see who is watching.

For balance, Frigate includes face recognition and licence plate recognition at no cost. Frigate+ is an optional paid subscription for fine-tuned detection models. The choice is therefore not a simple checklist of AI labels; it is a choice between Frigate's open-source, integration-led architecture and NetMagen Vision's native application and managed remote path.

Who should stay, and how to switch

Stay with Frigate if open-source code is a requirement, if Home Assistant or MQTT is central to the design, or if you enjoy maintaining Docker and YAML. Frigate is also a natural fit when a tuned Coral, Hailo or OpenVINO detector is already working as intended. Those are strong reasons to keep a system that matches its owner.

If the build work is getting between you and the recordings, test NetMagen Vision in parallel before changing anything. A small pilot makes the difference visible on your own hardware and with your own camera streams:

1. Install NetMagen Vision on the Windows, Ubuntu or supported Raspberry Pi computer you plan to keep running.
2. Discover selected cameras through ONVIF, let NetMagen Vision test their RTSP streams, choose a local drive and let the timeline fill.
3. Compare live viewing, event finding, detection markers, Telegram snapshots and clip export before moving the remaining cameras.

Keep the comparison practical. Find a motion event. Read a transcript line and jump to it. Draw a zone and watch the Telegram snapshot arrive. Restart the computer and confirm that recording resumes. If you want a recorder that works natively across Windows, Ubuntu and Raspberry Pi without Docker, YAML or required detector hardware, download NetMagen Vision and put your cameras on the timeline.