For smart-home integrators

Home Assistant maintenance at professional scale — no VPN, no on-site visits.

Manage all your customers' Home Assistant installations from one dashboard. Status at a glance, maintenance at the click of a button — and your customer stays in control at all times.

Hosted in Germany (Hetzner) GDPR by design 2 free instances
app.ha-fleet-manager.com / customers / house-zurichberg
HZ
Haus Zürichberg online
HA Yellow · Core 2025.11.2 · cus_h7q9rt
CPU
22%
RAM
2.4/ 4 GB
Disk
71%
Uptime 14d 06h
IP 10.0.20.14
Relay latency 38 ms
Last ping ↻ 6 s
142
Devices
1.847
Entities
38
Automations
6
Dashboards
Integrations · 24
mqtt zigbee2mqtt hue shelly sonos knx zwave_js +17
Critical logs · 24h 2 errors
[zigbee2mqtt] Adapter (deconz) disconnected — ECONNRESET 9h
[supervisor] Add-on "deconz" failed to start (exit 1) 9h
01 · CUSTOMER HA with plugin No open port, no VPN, no router settings.
02 · RELAY Cloud in Germany Status every 60 s. Tunnel only on request.
03 · INTEGRATOR Fleet dashboard Full HA browser access from any device.
The problem

Three sentences you've been explaining to your customers for years.

Every customer needs their own VPN setup.

WireGuard per customer, on-site router config, key rotation. Half an hour per installation, and that's just to get in.

On-site visits cost you half a day.

40 km out, "quick look", 40 km back. Fifteen minutes of work turns into four hours — eating your margin.

You hear about problems only when the customer calls.

Zigbee2MQTT has been crashing for 14 days, the heating automation hasn't run since Sunday. The customer notices Thursday evening. The call is inevitable.

Five reasons

What Fleet Manager does differently.

Each point closes a concrete gap in competing solutions. No marketing fluff — only what really matters day to day.

USP 01

The customer decides when you log in.

Every maintenance starts with a request inside the customer's HA — subject, reason, duration. The tunnel only opens after their click. Twelve hours max, then it closes automatically.

Unlike the always-on access of other remote tools.
USP 02

Real HA depth — not generic monitoring.

Integrations, HACS, add-ons, automations, dashboards, critical logs, version, host metrics — every 60 seconds. You see that Zigbee2MQTT has crashed before the customer notices.

Generic RMM tools only see "yet another server".
USP 03

Multi-tenant from day one.

One account, any number of end-customers, clean tenant isolation. Admin and technician roles. Your business scales without each new customer costing you linearly more time.

Existing remote-access tools are single-tenant — built for end-users.
USP 04

Made in Switzerland, hosted in Germany.

Hetzner Frankfurt, German UI, DACH data region. Developed in Switzerland, Swiss contract law. You never have to explain to your customer why their smart-home data lives in Oregon.

Competing platforms are predominantly US-centric.
USP 05

No new hardware required.

Works on any HAOS installation — Raspberry Pi, NUC, HA Yellow, HA Green, any x86 hardware. Install the plugin, paste the API key, done.

Other platforms only work with their own hardware.
WHAT THE CUSTOMER SEES
Repairs 1 pending task
Maintenance request from __INTEGRATOR__
Subject: "Z2M restart after update"
Duration: up to 2 hours
Generic preview — the final layout matches the Home Assistant frontend.
USP 02 · In depth

You see what's actually running — not just that it's running.

Generic monitoring tools measure uptime and CPU. We measure what actually causes pain in Home Assistant: stopped integrations, broken HACS plugins, an add-on stuck in a restart loop.

  • 147 integrations tracked — per installation
  • Installed add-ons visible per instance
  • CPU · RAM · Disk sampled every 60 seconds
  • Critical logs readable without SSH access
CPU
avg · 60s
28%
Uptime
since last boot
47d
Active integrations
147 · 24 vendors
147
mqtt zigbee2mqtt hue zwave_js spotify sonos +141
Critical log entries · last hour
12 · Z2M adapter unreachable
[zigbee2mqtt] Failed to connect to MQTT broker (3 retries)
[zwave_js] Driver crashed: ECONNRESET
[recorder] Database is locked
How does it work?

From plugin to maintenance — five steps.

No open port, no VPN at the customer's site, no router config. Everything originates from the customer — outbound only.

Connection HA plugin ↔ relay · persistent WebSocket, outbound only
Maintenance tunnel Integrator receives a temporary URL → opens HA directly in the browser
Lifetime Max. 12 h · automatic teardown · can be ended manually at any time
01 ·

Install the plugin

A custom integration from HACS at the customer site — no port, no VPN, no router config.

02 ·

Normal operation

Every 60 seconds a compact status (CPU, RAM, integrations, add-ons) is pushed to the relay.

03 ·

Maintenance request

From the dashboard, with subject, reason and duration. The customer sees it as an actionable item directly inside Home Assistant.

04 ·

Tunnel opens

You get a temporary URL and work in your browser on their HA — as if you were on-site.

05 ·

Automatic teardown

After the agreed time the tunnel closes. The customer can also end it at any time.

Who is it for?

For your first step.

FREE

You. Your first setup.

Individual · 1–2 HA instances

You run your own Home Assistant, maybe one for family or a second home. Try the fleet view — the Free plan currently covers two installations with the full feature set.

Start free

Coming soon

Another audience · TBD

Another persona will live here — we're still gathering which secondary audience resonates most.

Pricing

Honestly small, honestly clear.

We're in MVP. The Free plan is real, the Paid plan will come with final terms — no made-up numbers.

Free
0 € / month

For power users and starters. Full feature set within the limit.

  • Up to 2 installations
  • Full monitoring feature set
  • Customer-controlled maintenance tunnel
  • DACH hosting (Hetzner)
Start free

What happens after 2 installations? You'll be asked whether to upgrade to Paid. Nothing is billed automatically.

Trust & privacy

Made in DACH. Not as a sticker.

Our architecture is privacy-by-design — not because we have to, but because it matters between you and your customer.

Hetzner Frankfurt

Data stays in the DACH region — with the exception of selectable relay servers for the maintenance tunnel.

Consent toggle

No access without the customer's explicit approval. Tunnel time-limited automatically.

No third-party trackers

This website sets no cookies, loads no analytics, no pixel.

Ready?

Manage your HA customers centrally.

Two installations free, full feature set. No credit card. No commitment.

Start free now Questions? Drop us a line