@ShadowFist Great question — and to be clear up front, Reolink’s platform and storage model are genuinely solid. I ran their cameras and NVR for years, all integrated with Home Assistant. They met my expectations for a long time.
Where things eventually broke down for me was in a few specific workflows I could never quite get right even with HA in the mix. For context, my setup includes both access control (front gate + front door) and exterior cameras, so some of these needs are a little niche.
1. Triggering recording when I buzz someone in
One of my biggest use cases was:
“If I unlock the gate for someone, record exactly who I let in.”
I could never do that reliably with Reolink. I worked around it by relying on motion recording, but that only helped after the person walked into the camera’s field of view. For deliveries or unknown visitors, I really wanted:
- event‑triggered recording
- linked to the access event itself
I wasn’t able to achieve that in a clean, deterministic way.
2. Viewing cameras on any TV without friction
My Reolink NVR was in a closet in my den. I could:
- view cameras on the den TV (hardwired)
- use the phone/iPad/app
- use my HA dashboard
But if I was in another room watching TV and heard something outside, there was no convenient way to pull up the cameras on the TV I was actually in front of. I had to go find my phone or walk back to the den.
That friction mattered more than I realized.
3. Reviewing multi‑camera events took a lot of effort
In an incident that happened in my neighborhood, I discovered just how manual the workflow was:
- scrub each camera to the timestamp
- find the moment
- export each clip individually
Even with HA, stitching together an event across multiple cameras was a chore.
Why I eventually looked outside the Reolink ecosystem
You’re absolutely right that Reolink’s “local-first” approach is a huge advantage. I even replaced a Ring doorbell with a Reolink doorbell at one point specifically because I didn’t want cloud dependence.
But when I did a 5‑year review of my whole setup, I realized:
“My expectations haven’t evolved even though the technology has.”
That’s what pushed me to look at other architectures.
What the UniFi Protect + Access combo unlocked for me
I’m not trying to “sell” anything — just explaining what solved the three issues above.
1. Event‑based recording tied directly to access events
When I unlock the gate or door — or when a recognized face or delivery PIN code is used — I can tell UniFi:
“Record Camera A and Camera B for this event.”
And Protect automatically:
- records the relevant cameras
- syncs them in a unified event view
- links the event directly in the access logs
This solved my original problem and made the workflow comically easy compared to how I was doing it before.
2. Unified “Activity View” + AI-enriched metadata
I can set cameras to record continuously but only surface “activities” I care about:
- motion
- person detected
- known face detected
- animal
- audio events
- smoke alarm heard
And with the AI Key, each activity gets a natural‑language description of what was seen or heard. Searching for events becomes trivial.
This is something I couldn’t replicate with Reolink + HA.
3. Apple TV integration removed all friction
Because my home uses Apple TV on every screen, the Protect app has become the “household control surface” for camera visibility.
If something happens outside:
- click Protect on the Apple TV
- see all live feeds
- click one → jump straight to timeline of “activities”
This changed the way my household interacts with the system.
4. Doorbell + access control features I didn’t expect
Once I moved into the ecosystem, I realized I could do things I wasn’t even looking for at the time:
- dual‑camera doorbell (face + package cam)
- automatic recognition of known visitors
- delivery codes with time windows
- tied into the same event system as the cameras
These weren’t “must haves” originally — but they removed a lot of small frictions I didn’t know I had.
Full write‑up of the Protect/AI Key transition:
Closing thought
I started with Reolink back in 2016 and it served me very well. It’s reliable, local, and perfectly fine for many use cases. My evolution wasn’t about “Reolink is bad” — it was about asking:
“Does my experience still need to be like this?”
And once I reframed the problem, I realized the system could behave in a calmer, more integrated way.
