Marine aquarium automation systems:
platform dossiers, reliability & compatibility
A full market map of marine aquarium automation systems: controllers (commercial and DIY), their per-component reliability, community "pro / con" sentiment, and a compatibility matrix against every piece of peripheral equipment. The focus is on reliable solutions with a single interface and a closed-loop alkalinity (Alk) circuit. A separate dossier is compiled for each system.
How to read this report
1Methodology and the nature of the data
How the report was assembled and how far it can be trusted
The material is the result of a multi-wave parallel investigation (43 agents) prioritising primary sources: specialist forums (reef2reef, ultimatereef, the CoralVue forum), video reviews with multi-year observations (BRStv, Reef Builders), GitHub repositories to gauge the maturity of software integrations, official manufacturer documentation and retail catalogues.
2The architectural principle of reliability
A three-layer model and why the controller must not be the single point of failure
| Layer | Purpose | Representatives |
|---|---|---|
| 1. Controller (the brain) | Data collection, switching of outlets/pumps, executing logic | Apex, GHL, Hydros, reef-pi, ESPHome |
| 2. Specialised modules | KH auto-titration, dosing, auto top-off, pH/ORP cut-offs | Trident, KH Director, Alkatronic, dosers, ATO, Milwaukee |
| 3. Interface | Visualisation and control | Native app, Home Assistant + tablet, custom dashboard |
Life-critical loops — heating, top-off, dosing, return — should run on standalone hardware with hardware-level fail-safes. The controller and Home Assistant act as a layer of monitoring, visualisation and non-critical orchestration, but must not be the only link in the safety chain. As phrased in the thread "Controller — Single Point of Failure" (reef2reef): "Most have stopped using controllers for control — only for monitoring."
3Market map: every automation system
A full overview; below (section 4) each key system is broken down in dossier format
The market counts more than twenty platforms plus a separate class of auto-testers. Detailed dossiers are given for the five serious contenders for the target task; the rest are summarised in overview tables (section 4.6).
| Platform | Status | Reliability | Local API / HA | Closed-loop Alk | Tier |
|---|---|---|---|---|---|
| Neptune Apex A3 (Jr / A3 / Pro) | Shipping (three tiers) | High (2 defects) | Best — local R+W | Yes (Trident+DOS) | Top |
| GHL ProfiLux 4.1e | Shipping | Best in class | Local-first (optional myGHL cloud); no official API | Yes (KH Director) | Top |
| Hydros X10 + iV | Shipping | Good | Cloud, beta, read-only | Yes (all-in-one) | Top |
| AquaPi / ESPHome (DIY) | Active development | = build discipline | Native local HA | DIY (+Alkatronic) | Top (DIY) |
| Red Sea ReefBeat | Shipping (ReefControl — H2 2026, no confirmed date) | Mixed | ReefDose local; ReefSense no | No auto-KH | Niche |
| reef-pi | Maintained (fork) | Hobby-grade | REST + MQTT | No | Workable |
| Aquatronica | Shipping (EU) | Stagnant | No | No | Niche |
| Reef Factory | Bankrupt (05.2025) | Cloud at risk (still operational; acquisition talks) | No API | Yes (cloud, for now) | Avoid |
| Neptune ApexEL / A2 · Maxspect ICV6 | Discontinued | Legacy | Apex local / none | Via modules / none | Dead |
Do not exist (verified): GroBot, the "Sentinel" controller, "Ultimate Aquatics", "Mudra". Aquaforest does not make controllers. Dead crowdfunders: Mindstream, MOAI.
4System dossiers
For each: capabilities · "pro / con" opinions · compatibility with equipment and other systems
4.1 · Neptune Apex (A3 / A3 Pro series)
TopThe most complete ecosystem and the best dashboard; optimal if your flow is built on EcoTech Vortech and a ready-made Home Assistant integration matters.
Measurement
The A3 line has three tiers (A3 Jr / A3 / A3 Pro): the A3 Jr is monitoring-only (Wi-Fi only, 2 FMM, no EB832, sub-$300), the A3 adds Wi-Fi+Ethernet and 3 FMM (no EB832), and the A3 Pro has 4 FMM and includes the EB832. Temperature, pH, level and leak come via PM1/PM2/FMM modules; ORP is a built-in probe only on the A3 Pro (the A3 / A3 Jr need a PM1), and no A3 model ships a salinity probe — salinity requires a PM2 module plus probe across the whole line. Alk/Ca/Mg — the Trident auto-titrator.
Control
Dosing — DOS, including the current DOS QuietDrive ($374.99), which requires an A2- or A3-Series Apex (not the Apex Classic) and connects via 1Link to the EB832; flow — Vortech (MXM) and third-party pumps (VDM, 0–10V); lighting — EcoTech/AI (MXM); outlets with per-channel metering — EnergyBar EB832.
Reliability
High; two defects: "Cross-Chip" (outlets stick with no alarm) and EB832 PSU failure. The Trident is the most troublesome unit, but with a protective fall-back to a baseline dose.
Interface
Apex Fusion (cloud) is the benchmark. The local web UI over a static IP gives full monitoring and control without internet. The best local API.
Pro
- "Best software and support in the industry; Fusion is the interface benchmark"
- "4–5 years without a single problem"; there are owners of the original units
- The largest ecosystem and compatibility with third-party gear
Con
- "Went through 3 head-unit crashes in 2 years"
- "Cross-Chip": outlets lock in their last state with no notification
- EB832 PSU failure — a DIY fix exists (Harry's Aquatics repair/upgrade kit from $9.99, up to a pre-assembled Meanwell IRM-10-12 kit); Neptune out-of-warranty replacement is ~$149; US-centric warranty
itchannel/apex-ha)4.2 · GHL ProfiLux (4 / 4.1 / 4.1e)
TopThe benchmark for hardware reliability and fully local, cloud-free operation; optimal when the priority is longevity and locality, and the flow can use Tunze/Maxspect instead of Vortech.
Measurement
Temperature, pH, ORP, conductivity (on-board on the ProfiLux 4), level. Alk — KH Director; Ca/Mg/K/Na/NO₃ — the ION Director (five measured parameters, not just Ca/Mg).
Control
Dosing — GHL Doser; flow — Tunze/Maxspect/Abyzz (1–10V); lighting — Mitras; outlets — the Powerbar 6E-PAB (6 independently switched sockets up to 15A each, per-socket real-time current metering, PAB-bus daisy-chaining, status LEDs). Mobius-Vortech is not supported.
Reliability
Best in class (there are units running since 2012). The weak spot is the software layer and Wi-Fi dropouts on non-Ethernet models (get the 4.1e with LAN). The KH Director has finicky probe calibration.
Interface
Local-first: the built-in web server runs fully locally, with optional myGHL cloud remote access (ProfiLux 3/4, Doser 2.x, Mitras LX7); a graphic display on the case. No official API; HA is a workaround scrape.
Pro
- "GHL is the most reliable. Period"; "the ProfiLux 3.1 has been running since 2012"
- "Built like a tank"; runs fully local without the cloud
- Several operators switched from Apex to GHL for longevity
Con
- The thread "GHL ProfiLux 4… don't do it" (Reef Central)
- "Loses Wi-Fi, needs a full reboot" (non-Ethernet models)
- "An engineer's interface"; slow pages; "the lights come back on when you log out"
- The Doser 3 needs periodic lubrication of its planetary gears/rollers/tubing (food-safe silicone, sold by GHL — see the Doser 3 page) and has documented issues (heads stall on new tubing, daily Wi-Fi drops needing a power cycle, a dose-field firmware bug — see the Doser 3 disappointment thread)
4.3 · CoralVue Hydros (X10 + iV)
Top · all-in-oneA controller and four stepper dosers in one case; the closed-loop alkalinity titration is added via a separate iV module (sold separately; in use it occupies 3 of the 4 doser channels). The most compact route to auto-Alk; the best fail-safe dosing settings.
Measurement
Temperature, pH, ORP, salinity. Closed-loop Alk — a separate iV module (sold separately; occupies 3 of the 4 dosers).
Control
4 built-in dosers; flow — WaveEngine (4 mixed-brand pumps over 0–10V). Vortech is not supported.
Reliability
The wired controllers are sturdier than the Apex base. Caveats: the Wi-Fi bar is unreliable (get the XP8); the iV needs lengthy setup; pH/salinity probes drift.
Interface
A quality app, but control runs through the cloud (you cannot change settings without internet). HA is beta, monitoring only.
Pro
- "Build quality is sturdier than Apex. Nothing failed in 2 years"
- "The XP8 works flawlessly" — the best power bar by consensus
- Distributed architecture — no single point of failure; wired modules keep working when the internet is down
Con
- "The Wi-Fi bars are pure junk, do not put anything important on them"
- "Over a month of back-and-forth with CoralVue before the iV setup worked"
- The thread "Have Hydros lost their way?" (pricing drift: Hydros Sole 1-head $346 vs Neptune DOS QD / GHL Doser 2.2 Maxi 2-head ~$450/~$369; Hydros Duet 2-head $800 — a free Reef2Reef account is required to view it); probe drift, early RMAs
Bitf1ip/ha-hydros)4.4 · Red Sea ReefBeat
NicheA single app for its own ecosystem; there is no native auto-KH, so a closed-loop alkalinity circuit requires a third-party titrator. The ReefDose doser has the best local HA integration among dosers.
Ecosystem
ReefLED lighting, ReefWave pumps, the ReefDose doser, the ReefATO+ top-off, the ReefRun DC return/sump pump, the ReefMat roller — all under one app. The ReefControl controller has not shipped yet (H2 2026, no confirmed date, no real reviews) and ships in tiers — ReefControl Lite (~$329, up to 2 probes), Pro (~$429, up to 4 probes, expandable to 7) and a Pro Deluxe Kit (~$699, 3 probes + leak sensor). The ReefSense probe set is pH/Temp, Salinity/Temp, ORP and Temp only — no KH/alkalinity probe (verified June 2026; SaltwaterAquarium.com shows an est. ship ~Aug 10 2026).
Closed-loop Alk
Not possible natively (no auto-titrator). Solved by adding a standalone Alkatronic.
Reliability
ReefDose is highly rated; the ReefWave cable hardens and splits at ~3 years (one user lost 3 of 4 ReefWave 45 units; a free Reef2Reef account is required to view the thread). The app is the most criticised component.
Interface
The native app is weak; ReefDose has the best local HA integration among dosers (read and write).
Pro
- "ReefDose and ReefMat are fantastic, very well designed"
- "Over a year — not a single problem with the app or Wi-Fi" (some owners)
Con
- "The app wipes the whole tank setup almost every day"
- The thread "NEVER, EVER BUY A RED SEA ANYTHING" (June 2026)
- ReefWave cable degradation; ReefATO+ overflow incidents
Elwinmage/ha-reefbeat-component) — best among dosers; the same repo also covers the ReefRun DC return/sump pump (Tested)4.5 · DIY: Home Assistant + ESPHome / AquaPi
Top for the "one tablet" priorityA fully local, brand-agnostic layer: it brings any equipment into a single tablet panel. On-device logic (ESP32) survives a Home Assistant outage. The best fit for the single-screen target task.
Capabilities
Measurement of all parameters (ESP32 + Atlas EZO digital probes for pH/ORP/salinity), relay/outlet switching (Shelly), lighting, dosing. Closed-loop Alk — an external standalone titrator (Alkatronic).
Interface
One tablet with all the information from every brand in a single local panel; a custom dashboard (HA WebSocket API) or Lovelace + Fully Kiosk.
Reliability
Determined by build discipline and the presence of hardware fail-safes. No cloud, no subscriptions.
Ready-made projects
AquaPi (ESP32/ESPHome, kits are sold), reef-pi (REST+MQTT, but finicky), ReefRhythm (telenkov88/reefrhythm-smartdoser, an active ESP32-S3 / MicroPython stepper doser, 19 releases, MQTT since v0.8) / DIY my Dose (dosing).
Pro
- "On-device ESP32 logic survives an HA outage — the tank doesn't go down with the server"
- Fully local, no cloud; a single dashboard for any brand
- AquaPi is actively developed; ESPHome is the fastest-growing segment
Con
- "reef-pi nearly cooked the tank" — with a failed sensor the heater was left on
- Reliability of critical loops is "a compromise", hardware backups are needed
- You build and maintain it yourself; no ready-made closed-loop Alk
jrigling; other models — cloud) · ⚠️ Vortech (Feed only, BLE bridge)4.6 · Focustronic (Alkatronic · Dosetronic · Mastertronic)
Top for auto-AlkNot an aquarium controller but a standalone ecosystem for closed-loop chemistry maintenance — the one people most often switch to from the Trident and KH Director. It sits next to any brain and does not require one.
Measurement
Alkatronic — KH titration (pH electrode), 1–12 tests/day; the original retains a BNC output, while the Alkatronic Pro (2024, $899.99 / €1,225) adds pH+temp monitoring, Powertronic heater/chiller switching, Dosetronic kalkwasser control, voice and a DC pump but removes the BNC output. Mastertronic — multi-test (KH/Ca/Mg/NO₃/PO₄).
Control
A closed loop: the Alkatronic corrects alkalinity with its own pump or commands a Dosetronic doser (5 channels), including proportional Ca/Mg correction. The new Tracetronic precision trace-element doser (announced Q4 2025, page now live) doses up to 12 trace elements from 12×40 ml vials via a single-hose patented delivery system at 0.1 ml accuracy, and can read ICP lab recommendations to auto-adjust dosing with user authorization (no price published at announcement).
Reliability
The best long-term reputation among auto-Alk; the cheapest reagent (~$0.04/test, ~1 year). Weak points — SD-card corruption from an outdated app and an auto-mode catastrophe. The Mastertronic Essential is raw.
Interface
Its own app (the dosing loop runs locally on the device). In Home Assistant — cloud scrape only, read.
Pro
- "0.01 dKH variance over 2 years"; "the single biggest contributor to the tank looking good"
- Commercial coral farms run them for years; the cheapest reagent, reagent is made yourself
- Fully standalone — works next to any controller
Con
- "Alkatronic mode overdose": the dose spiked to 10,554 ml/day, fish died
- SD-card corruption when using an outdated app; noisy operation
- Harsh reviews of support; the Mastertronic Essential is "practically unusable"; HA cloud-only
stvnmbr1/hassio_focustronic)4.7 · Kamoer (KH Carer · F4 Pro · X-series)
Workable · mid-budgetA mid-tier dosing and auto-KH ecosystem. The pumps are precise; the software layer is a historical weak point. The only ecosystem with a documented cross-brand auto-Alk bridge (into Hydros).
Measurement
KH Carer — alkalinity titration (with its own pH probe). Reef Master SPA — multi-test (raw, not recommended).
Control
The KH Carer + the E1 expansion block drives F4 Pro dosers (closed loop). Dosers: F4 Pro V2 (4 channels), X1 Pro, FX-STP (continuous, controlled by an outlet).
Reliability
Precise stepper heads; reagent is made yourself (~$20/year). Risks: clock drift offline, cloud outages, false drop-sensor errors, "bricking" on a firmware failure, a KH Carer error of ~2.5 dKH on a reagent batch change.
Interface
The KamoerRemote app (cloud). There is no working local HA integration.
Pro
- "After calibration the KH Carer matches the Hanna"; a precise stepper drive
- The best price/reliability ratio in the mid segment; DIY reagent (~$20/year)
- The only auto-Alk with an official cross-bridge — feeds readings into Hydros
Con
- "The KH Carer read ~2.5 dKH low" after a reagent batch change — dangerous in closed-loop dosing
- Clock drift offline; cloud outages block access; false drop-sensor errors
- Confirmed "bricking" on an incomplete firmware update; software is the weak link
4.8 · D-D (KH Manager · P4 Pro)
Workable · the EU/UK Kamoer routeA joint D-D and Kamoer development: the same hardware as the KH Carer, but tuned and supported by D-D for the EU/UK market. The right way to buy this platform in Europe — with a local warranty instead of importing through the US.
Measurement / control
KH Manager — alkalinity titration; with the E1 block it automatically corrects up to two P4/P4 Pro dosers (multi-element correction). The E1 can also drive a single Kamoer FX-STP to control supply-water flow to a calcium reactor (a CaRx loop). Auto-retest of outlier readings before dosing; user dose limits.
Reliability / interface
Early reviews — an exact match with the Hanna after setup; non-proprietary reagent. Inherits the platform's risks (drop sensor, firmware). Long-term data is still scarce. Cloud for remote/HA, but a Bluetooth local-connection mode also exists (no local HA integration).
Pro
- "Matches the Hanna Marine Master exactly after setup, stable for 4 weeks"
- Non-proprietary reagent; auto-retest before a dose; user dose limits
- Best EU/UK support and warranty for the Kamoer platform
Con
- The same hardware platform as the KH Carer — the same drop-sensor and firmware risks
- Few long-term reviews (a relatively new product)
- The automation depends on the cloud; no local HA integration
4.9 · Aquatronica
NicheAn Italian full-featured controller with a 7" touchscreen, EU stock and 230V. The hardware is durable, but the platform is effectively stagnant: no API, no auto-KH, fragile firmware updates.
Measurement / control
Temperature, pH, ORP, conductivity, level; dosing (by timer or pH/ORP trigger); outlet switching; lighting. No KH auto-titrator — a closed-loop alkalinity circuit is impossible.
Reliability / interface
There are ~15-year cases on original components, but PC firmware updates are documented to "drop" outlets/dosing; English-language support is weak. A 7" touch + app; no local API or HA.
Pro
- "~15 years mostly on original parts"; mature hardware
- EU-native, 230V Schuko, available in European shops
Con
- "A firmware update dropped the outlet/dosing setup"; the UI reverted to Italian
- The platform is stagnant (hardware 2011–2014), no new development
- No API/HA, no auto-KH, a thin English-language community
4.10 · Tunze (pumps · Osmolator · SmartController)
The benchmark for component reliabilityNot a brain for the whole system but a set of the most reliable components: flow pumps with an open 0–10V drop into any controller, and the Osmolator is the best top-off. The SmartController is closed but remains a current niche product (limited retailer stock, no new firmware) — people mainly buy the components.
Lineup
Turbelle pumps (open analogue 0–10V input), the Osmolator 3 (model 3154) top-off (optical control sensor + an independent thermal safety sensor, no float, with alarm), and the SmartController 7000 (a current niche product, limited retailer stock, no new firmware development).
Reliability / integration
The best longevity in the category (pumps lasting 6–9 years), no cloud. The pumps integrate into Apex (VDM), GHL (PLM), Hydros (WaveEngine), Home Assistant (via ESP/0–10V). The Osmolator runs standalone, with its alarm wired into a controller input.
Pro
- "The 6105 has run ~9 years without a hitch"; "practically indestructible", "absolutely silent"
- Open 0–10V — no cloud ecosystem; drops into any controller
- The Osmolator is the most reliable ATO (optical + thermal protection, no float)
Con
- Bulky form factor; a utilitarian app
- The SmartController is a closed niche product, no API, no new firmware
- No auto-KH — an external titrator is needed
4.11 · Jebao / Jecod
Budget · the only one with local HAThe ubiquitous budget ecosystem for flow, return and dosing. The paradox: despite being cheap, this is the only budget brand with true local control in Home Assistant (no cloud) — yet its dosers have the most dangerous reputation on the market.
Lineup
Wavemakers (MOW / OW / SLW), return pumps (DCP / DCS / MDP), dosers (DP-4 / MD-4.4). A Gizwits cloud chip; the new SW-series wavemakers have a 0–10V input for DIY.
Control
Wi-Fi via the Jebao Aqua app; locally — through Home Assistant. There is no metered external control of the dosers.
Reliability
The wavemakers are decent (2–4 years). The dosers are bimodal: usually fine, occasionally they "dump the whole bottle". No warranty. New revisions (ESP8266 → ESP32-C3) changed the protocol and are less predictable.
Interface
A basic cloud app. The local HA integration (jrigling) works on an isolated IoT network without internet.
Pro
- "The DP-4 has run 5 years, great value" (some owners)
- SLW wavemakers run quietly for 2–4 years; the cheapest entry
- The only budget brand with true local control in HA (jrigling), no cloud
Con
- "The DP-4 dumped the whole bottle" of Alk overnight — tank destroyed (the classic complaint)
- "Jebao gives no warranty at all — it's all on the shop"
- New revisions are less reliable; clock drift; lines airlock; fail-safes are needed
jrigling) — only the MDP-20000 is actually confirmed; SW/DC wavemakers — cloud (chrisc123). The new SW wavemakers have a 0–10V input for DIYjrigling — local (young, a narrow model list) · ⚠️ chrisc123 — writes via the cloud, breaks on new hardware4.12 · Other systems (overview)
| System | What it is · key "pro / con" | Tier |
|---|---|---|
| reef-pi (DIY) | An open Pi controller, full REST+MQTT. Pro: openness, HA; still gets substantive commits (June 2026). Con: "nearly cooked the tank"; official releases stalled since v6.0 (Nov 2022); broken on Pi OS Bookworm and newer (libgpiod/GPIO changes; Bullseye 32-bit is the workaround; compat issue #2194 was closed "not planned"); an experimental community fork targets Trixie 64-bit (repo). | Workable |
| Aquatronica | An Italian controller with a 7" touch. Pro: the hardware lasts, EU stock. Con: stagnant, no API, no auto-KH, fragile firmware updates. | Niche |
| Tunze SmartController 7000 | German, mature, a current niche product (limited retailer stock, no new firmware). Pro: Tunze build. Con: closed, no API, no auto-KH. | Niche |
| Aqua Medic | A set of Wi-Fi devices (essentially Jebao on the Gizwits cloud), not a single controller. Pro: DC Runner x.3 pumps have a confirmed community HACS integration (maziggy/aqua_medic_dc_runner, via Gizwits). Con: EVO calibration drift, cloud; the Reefdoser EVO is only in beta/in-development (Elwinmage/ha-aquamedic-component), not yet confirmed in HA. | Components |
| Kamoer (X-series, KH Carer) | Dosers + auto-KH. Pro: a precise stepper drive, price. Con: weak software, cloud, a KH Carer error of ~2.5 dKH on a reagent batch change. | Workable |
| Pacific Sun Kore 7th | A controller with a built-in KH lab. Con: little-known, few reviews, no API documentation. | Niche |
| Robo-Tank / Oceanberry / Marine Assistant / Reef Angel | DIY controllers (ready-made or kits). Pro: openness, locality. Con: varying maturity; Reef Angel is legacy, the "Star" is buggy. | DIY |
| Reef Factory Smart Reef | A Polish cloud ecosystem. Con: bankruptcy (05.2025) — its cloud and 360 Smart Reef app reportedly remain operational (distributors underwriting hosting) with acquisition talks ongoing; a Level Keeper flood (~€5000), doesn't work without internet, refused an API. | Avoid |
5Compatibility by equipment type
A cross-cutting view: one piece of equipment — all controllers (complements the dossiers above)
5.1 · Flow pumps
| Pump | Input type | Apex | GHL | Hydros | HA / DIY | Longevity |
|---|---|---|---|---|---|---|
| Tunze Turbelle | 0–10V OPEN | ✅ VDM | ✅ | ✅ WaveEngine | ✅ ESP+DAC | benchmark, 8–10 yr (10+ with care; PSU ~yr 4) |
| Reef Octopus Octo Pulse / VarioS | 0–10V OPEN (native) | ✅ | ✅ | ✅ | ✅ ESP+DAC | good, underrated |
| Abyzz | 0–10V OPEN (interface) | ✅ | ✅ | ✅ | ✅ | premium, 10 years |
| IceCap Gyre 2K–4K | 0–10V OPEN (via WaveEngine LE controller) | ✅ | ✅ | ✅ WaveEngine LE bundle | ⚠️ 0–10V | gyre drive wear |
| Maxspect Gyre | 0–10V via IceCap adapter | ✅ | ✅ | ✅ WaveEngine | ⚠️ | controller failures |
| EcoTech Vortech MP40 QD | Mobius BLE CLOSED | ✅ MXM | ❌ | ❌ | ⚠️ Feed only | high (wet side is a consumable) |
| AI Nero 5 | Mobius BLE CLOSED | ✅ MXM | ❌ | ❌ | ⚠️ | variable QC |
| Sicce XStream SDC / Syncra SDC | ContrALL Wi-Fi CLOSED | ⚠️ Apex Fusion only | ❌ | ⚠️ wireless slave | ❌ | 3-yr base + 2-yr via registration (=5); impeller 6 mo |
| Sicce XStream-E | 0–10V OPEN (RJ45) | ✅ VDM | ✅ | ✅ WaveEngine | ✅ ESP+DAC (control cable) | high, 5-year warranty |
| Coral Box QP | 0–10V via AquaLink A1 | ✅ adapter | ⚠️ | ✅ WaveEngine | ⚠️ | variable QC |
| Jebao SW / DC / Wi-Fi | 0–10V partial / Wi-Fi | ❌ | ❌ | ⚠️ DC via WaveEngine | ⚠️ MDP local; SW/DC — cloud | controller 1–2 years |
| Red Sea ReefWave | ReefBeat cloud CLOSED | ❌ | ❌ | ❌ | ⚠️ ha-reefbeat-component | noise, controller failure |
| Waveline DC | dumb remote | ❌ | ❌ | ❌ | ❌ | poor — avoid |
Pump connections. OPEN (0–10V): Tunze, Reef Octopus, Abyzz, IceCap Gyre — drop into any controller and directly into Home Assistant (ESP32 + GP8403 DAC, local). CLOSED: Vortech/Nero — Apex only, via MXM; Sicce SDC — Apex Fusion / wireless slave only, but the XStream-E is OPEN (0–10V via RJ45); Red Sea ReefWave — its own cloud only. Longevity ranking: Tunze > Abyzz > Sicce > Vortech > Reef Octopus > Gyre > Nero > ReefWave > Jebao > Waveline (the AI Nero is downgraded for frequent early failures in 2024–25). Note: a "Neptune WaveEngine" does not exist — WaveEngine is a Hydros product; Neptune's equivalent is the VDM module + 0–10V ports.
5.2 · Return pumps
| Pump | Input type | Apex | GHL | Hydros | HA | Standalone Feed | Longevity |
|---|---|---|---|---|---|---|---|
| Royal Exclusiv Red Dragon X | no external input | outlet on/off | on/off | on/off | smart outlet | yes (F button on the unit) | top tier, 10+ years |
| Neptune COR-20 | AquaBus (Apex-only) | ✅ native | ❌ | ❌ | ⚠️ apex-ha | yes (Feed button) | quiet, reliable |
| Neptune COR-15 | 1Link (Apex-only) | ✅ native | ❌ | ❌ | ⚠️ apex-ha | no (requires Apex) | quiet, reliable |
| Red Sea ReefRun | ReefBeat CLOSED | ❌ | ❌ | ❌ | ✅ local (ha-reefbeat-component) | via app | new, little data |
| Abyzz | 0–10V OPEN | ✅ | ✅ | ✅ | ✅ ESP | — | premium, 10 years |
| Tunze Silence / Recirc (controllable) | 0–10V OPEN | ✅ VDM | ✅ | ✅ | ✅ ESP | — | benchmark |
| Reef Octopus VarioS (return) | 0–10V OPEN | ✅ | ✅ | ✅ | ⚠️ | — | good |
| EcoTech Vectra | Mobius BLE CLOSED | ✅ MXM | ❌ | ❌ | ⚠️ Feed | via app | Gen2 ok, weak PSU |
| Sicce Syncra SDC | ContrALL Wi-Fi CLOSED | ⚠️ Apex Fusion | ❌ | ⚠️ wireless slave | ❌ | — | high, 5-year warranty |
| Jebao DCP / DCS / MDP | Wi-Fi / 0–10V partial | ❌ | ❌ | ⚠️ | ⚠️ MDP-20000 only (jrigling) | via app | weak controller |
Return connections and the Feed button. The return is the least "smart" part; when feeding, it is often left untouched (only the wavemakers are paused). The key nuance: the Red Dragon X has no external control input — an external Feed is only possible by cutting outlet power, and then siphon-break protection is mandatory (a hole at the water line). The COR-20 (AquaBus, ships with a 3-ft cable) and Vectra can do Feed, but the Vectra is via Mobius. The COR-15 requires an Apex (1Link) and has no standalone mode (BRS redirects non-Apex buyers to the COR-20). For unified closed-loop return control with local HA, use an OPEN pump (Tunze / Abyzz / Reef Octopus VarioS, 0–10V) or Jebao (local). Premium reliability without control — a Red Dragon X on a controlled outlet + siphon protection.
5.3 · Lighting
| Light | Control type | Apex | GHL | Hydros | HA | Reliability |
|---|---|---|---|---|---|---|
| Philips CoralCare Gen2 | 0–10V OPEN (+ own controller) | ✅ | ✅ | ✅ | ✅ Shelly 0–10V | top tier (metal, 24-month warranty) |
| Kessil A-series / AP9X | 0–10V OPEN + app | ✅ native (IOTA) | ✅ | ✅ | ✅ ESP/Shelly | mixed-to-good |
| Reef Brite XHO / Lumi | 0–10V OPEN | ✅ VDM | ⚠️ 0–10V | ✅ | ✅ Shelly | durable LEDs (weak BT remote) |
| Red Sea ReefLED | ReefBeat (Wi-Fi) | ❌ | ❌ | ❌ | ✅ local (ha-reefbeat-component) | good LED, weak app |
| EcoTech Radion | Mobius BLE CLOSED | ✅ MXM | ❌ | ❌ | ⚠️ DIY (Feed) | top optics, Mobius software is finicky |
| AI Hydra 64HD | Mobius BLE CLOSED | ✅ MXM | ❌ | ❌ | ⚠️ HydraBridge ESP32 (local, 64HD only) (repo) | good thermals, Mobius software |
| AI Prime / Hydra (other) | Mobius BLE CLOSED | ✅ MXM | ❌ | ❌ | ❌ no HydraBridge | good thermals, Mobius software |
| GHL Mitras LX7 | proprietary digital (ProfiLux) | ❌ | ✅ native (9 ch.) | ❌ | ❌ | stable on ProfiLux |
| Orphek / Maxspect / Zetlight | cloud / proprietary CLOSED | ❌ | ❌ | ❌ | ❌ | varies |
| Noopsyche K7 V3 / Mini / Pro IV | Wi-Fi (locally hacked) | ❌ | ❌ | ❌ | ⚠️ unofficial ESP32-S3 (or PC/Android) browser web UI, drag-and-drop scheduler | fan noise/failure |
Lighting connections. The most reliable local route — an OPEN fixture (Philips CoralCare or Kessil) + a Shelly 0–10V dimmer in Home Assistant: everything runs locally, no cloud and no vendor app. The Red Sea ReefLED is the best of the "closed" ones, because ha-reefbeat-component controls it over a local API. EcoTech/AI — only via Apex MXM (the AI Hydra 64HD has a local HydraBridge ESP32 bridge; the AI Prime does not). Orphek, Maxspect, Zetlight are closed cloud islands, no good for local HA.
5.4 · Skimmers (DC)
| Skimmer / DC pump | Pump control | Apex | GHL | Hydros | HA | Reliability |
|---|---|---|---|---|---|---|
| Reef Octopus VarioS (Regal/Diablo) | 0–10V OPEN (native) | ✅ VDM | ✅ | ✅ | ⚠️ 0–10V | best controllable, durable |
| Royal Exclusiv Red Dragon | 0–10V via adapter (~$211) | ✅ | ✅ | ⚠️ | ⚠️ | best wet side; fragile controller |
| Deltec E-Flow | 0–10V OPEN | ✅ | ✅ | ⚠️ | ⚠️ | excellent build |
| Nyos Quantum (original) | manual AC knob (no smart control) | ❌ | ❌ | ❌ | ❌ | little data |
| Nyos Quantum EQ (160/220/300, Apr 2025) | Mobius CLOSED | ⚠️ Apex via MXM Module (sold separately) | ❌ | ❌ | ❌ | little data; an EQ upgrade kit retrofits existing bodies (details) |
| Red Sea ReefSkimmer DC | ReefBeat CLOSED | ❌ | ❌ | ❌ | ✅ ha-reefbeat-component | self-levelling is finicky; 30 % work without the controller |
| AquaMaxx / Skimz / Bubble Magus DC | own remote | outlet on/off | on/off | on/off | smart outlet | budget DC is unreliable |
Only the Reef Octopus VarioS, the Royal Exclusiv Red Dragon and the Deltec E-Flow are genuinely controllable over 0–10V (speed + feed pause). The other "DC" units only have their own remote with presets — an external controller can only switch them on/off. Red Sea's "self-levelling" is throttling by a neck-humidity sensor, not sump-level tracking; by all accounts it is finicky. The best controllable skimmer for automation is the VarioS; the most reliable wet side is the Red Dragon (an adapter is required).
5.5 · Roller / fleece filters
| Filter | Type | Sensor | Controllers | HA | Reliability |
|---|---|---|---|---|---|
| Red Sea ReefMat | Smart (advances by clog) | Dual conductivity (no optics) | none natively | ✅ local (ha-reefbeat-component): advance, roll remaining, alarms | the motor is the weak link (3–15 months) |
| Clarisea SK-5000 Gen3 | Smart | Float | DIY (motor + optics on Apex/GHL) | via DIY | a history of motor wear; repairable |
| Theiling Rollermat | Sensor-driven | Float | none | none | robust mechanics, little data |
| IceCap RollBot | Timer-driven | none | none | none | reliable (no sensor), but doesn't react to clogs |
| Reef Factory Smart Roller | Smart, cloud | Level | none | none | cloud + bankrupt → avoid |
The only roller with a true local HA integration is the Red Sea ReefMat (via ha-reefbeat-component, no cloud): advance on demand/schedule, roll remaining, clog/break alarms. The catch is the motor as a wear part: buy from a seller with a good RMA and keep a spare motor. The vendor-cloud-free alternative is a Clarisea Gen3 wired to your own controller. A "Nyos roller" does not exist.
5.6 · Reactors
| Type | Method | Apex | GHL | Hydros | Standalone |
|---|---|---|---|---|---|
| Calcium (CaRx) | pH probe → CO₂ solenoid, hold ~6.5 | ✅ programmable | ✅ native mode | ✅ DC solenoid | Milwaukee MC122 |
| Media (GFO/carbon) | Pump on an outlet | ✅ | ✅ | ✅ | any timer |
| Biopellet | A dedicated 24/7 pump | ✅ | ✅ | ✅ | — |
| Kalkwasser | Top-off/doser + high-pH cut-off | ✅ | ✅ | ✅ | Milwaukee MC122 |
Critical failures. Calcium: "CO₂ runaway" with a stuck solenoid → a pH crash. Mitigations: a two-stage regulator, two solenoids in series, a low-pH cut-off, a time limit. Kalk: an overdose → a pH spike and CaCO₃ precipitation; you need a high-pH cut-off (~8.4) + max-runtime. "Smart" media reactors do not exist.
5.7 · Algae scrubbers (ATS) and refugium
| Equipment | Light control | Native integration |
|---|---|---|
| Santa Monica, IceCap ATS, Clearwater / Adaptive Reef | AC outlet + an external timer | ❌ (Adaptive Reef plainly: "Apex Compatible: No") |
| Turbo's Aquatics (Rev 4) | A 3-channel dimmable driver | ✅ the only ATS with controller control (Apex/PWM) |
| Refugium: AI Blade Fuge / Prime Fuge | Own schedule (app) | ✅ + Apex via MXM |
| Refugium: Kessil H80, Neptune GRO | 0–10V / on-off | ✅ GHL/Hydros natively; Apex via VDM |
All four controllers can schedule a reverse photoperiod (light out of phase with the display). Harvesting is not automated. The circulation pump runs 24/7 — the single point of failure; per-channel current monitoring is recommended.
5.8 · Auto top-off (ATO)
2026 market corrections: the Tunze 3155 (with a float) is discontinued — the current Osmolator 3 = model 3154 (optical + thermal protection, no float). The AutoAqua Smart ATO Micro has one optical sensor (not two), and is winding down in the US. "Reef Octopus AWC", "Ghost AWC", "IceCap AWC" do not exist.
| Device | Sensors / redundancy | Standalone / controller | HA | Reliability (overflow risk) |
|---|---|---|---|---|
| Neptune ATK V2 | 2 optical + timeout + a mechanical float valve + alarms | Best with Apex | native (Apex) | Highest (4 layers) |
| Avast Marine Parrot | 2 pressure sensors + a float valve (nothing electronic in the water) | Apex/GHL/Hydros or standalone | via a controller | Highest (triple analogue) |
| Hydros Triple Optical | 3 optical (on / stop / independent protection) | Hydros only | beta | High |
| GHL ProfiLux ATO | optical/float + a time limit + leak | ProfiLux only | DIY | High |
| Tunze Osmolator 3 (3154) | optical + thermal protection (no float) | Standalone | smart outlet | Medium (optics fouling → sticking) |
| Kamoer ATO One 2 | up to 3 optical + a float + an alarm | Standalone (Wi-Fi) | LocalTuya | Medium (some glitches, the float saves it) |
| XP Aqua Duetto 2 | 2 optical in one body + timeout | Standalone | — | Medium (the worst failure/support history) |
| AutoAqua Smart ATO Micro | 1 optical + timeout (no float) | Standalone | — | Low (a single sensor) |
| Red Sea ReefATO+ | A 5-probe titanium EC array (incl. backup pump-OFF probes ~2.5 cm above shutoff) + a separate titanium leak detector | Standalone (ReefBeat) | local (ha-reefbeat-component) | The worst overflow history |
ATO ranking: Neptune ATK V2 (with Apex) ≈ Avast Parrot > Hydros / GHL (multi-sensor + timeout) > Tunze Osmolator 3 > Kamoer ATO One 2 > XP Aqua > single-sensor units (AutoAqua Micro) > Red Sea ReefATO+ (best HA, worst redundancy). A cheap mechanical float valve (Eshopps) — as a third emergency layer on any of them.
5.9 · Automatic water change (AWC)
| System | Type | Safety | HA | Reliability |
|---|---|---|---|---|
| Neptune DOS (Apex) | Controller-based, volumetric | I/O floats + alarms | strong (Apex) | Precision benchmark; recalibrate ~every 2 weeks |
| Hydros AWC (3 pumps) | Controller-based, by level | Leak input + timers + state memory across a power loss | beta | Best fail-safe behaviour |
| GHL ProfiLux AWC | Controller-based, by level | 2 sensors + max-on-time | DIY | Robust; the Maxi runs standalone |
| Kamoer X2SR / Pro | A dedicated Wi-Fi pump (integrated 24V DC ATO — one cord drives both AWC and ATO) (spotlight) | 4 overflow floats + alarms in the app | Tuya (unverified) | "Built like a tank"; the pain is Wi-Fi setup |
| AutoAqua Smart AWC Duo G2 | Dedicated, all-in-one (AWC + built-in ATO) | Dual optical + a QST alarm (local); the built-in ATO uses a patented dual optical sensor and is auto-disabled during a change and for 3 min after | — | Cheap, but QC complaints |
| Reefloat AWC | Dedicated, via a vessel | Float/level + cut-offs | — | Reliable, gentle; UK sourcing |
| Genesis Renew Pro | Dedicated, metered vessels | 2 sensors + a low-stock cut-off | — | Proven, "industrial"; expensive, backordered |
The Neptune ATK is an ATO, not an AWC. The main AWC danger: the ATO fires during a change and tops off with fresh water → a salinity crash; good systems (DOS, Hydros, GHL) disable the ATO for the duration of the change. Remote monitoring is critical.
5.10 · Dosers — a brand comparison
The doser is the actuator of the Alk loop. Key criteria: reliability, a safe failure mode, a native pairing with auto-KH (closed loop), and a local HA integration.
| Doser | Reliability | Closed-loop Alk with | Local HA | Brief "pro / con" |
|---|---|---|---|---|
| GHL Doser 2.2 | High, repairable | GHL KH Director | DIY scrape | Pro: "5+ years, incredibly precise". Con: Wi-Fi |
| GHL Doser 3 | Good, but documented issues | GHL KH Director | DIY scrape | Pro: precise, LAN/Wi-Fi/USB. Con: needs periodic lubrication; heads stall on new tubing, daily Wi-Fi drops, a dose-field firmware bug (thread) |
| Neptune DOS / QuietDrive | High, safe failure | Trident (needs Apex) | via apex-ha | Pro: "3 years problem-free", skips a dose rather than overflowing. Con: the original DOS is loud; both need an Apex (the DOS QuietDrive $374.99 requires an A2/A3, 1Link to the EB832) |
| Red Sea ReefDose 4 | High, ±0.5 % | no native auto-KH | best (read+write) | Pro: "precise, quiet, flawless", the best HA panel. Con: Wi-Fi setup lottery |
| Focustronic Dosetronic | Good hardware, weaker software | Alkatronic | cloud | Pro: "4 years of hard use". Con: support complaints, a rare ~1 L dump |
| Kamoer F4 Pro V2 | Precise hardware, weak software | KH Carer | cloud | Pro: precision, price. Con: clock drift offline, cloud outages |
| D-D P4 Pro | = Kamoer, EU support | D-D KH Manager | cloud | Pro: EU warranty. Con: the same Kamoer platform |
| Kamoer FX-STP | A continuous pump | — (controlled by an outlet) | via an HA outlet | Pro: the only one driven by an ordinary outlet (CaRx, AWC, kalk). Con: doesn't dose in portions itself |
| Bubble Magus T11 | Budget, more reliable than Jebao | — | none (an analogue island) | Pro: "bulletproof" for the money, an error alarm. Con: no Wi-Fi/API — not externally controllable |
| Jebao DP-4 | Bimodal (dump risk) | — | jrigling (partial) | Pro: cheap. Con: "dumped the whole bottle", no warranty |
| Jebao MD-4.4 | new Wi-Fi 4-ch | — | jrigling (planned, May 2026) | Pro: set to be the first fully local (no-internet) HA Jebao doser. Con: not yet released in the integration; distinct from the legacy DP-4 / DS-4 line |
| EcoTech Versa | Low — avoid | — | none | Con: fitting leaks out of the box, tubing degradation; "done with the Versa" |
Best "doser + auto-KH" pairs for a closed loop: GHL Doser + KH Director (one ecosystem, local) · Focustronic Dosetronic + Alkatronic (best standalone reliability) · Neptune DOS + Trident (best local HA, but needs Apex) · Kamoer/D-D — budget, but cloud. For a pure dosing HA panel outside a closed loop — the Red Sea ReefDose 4.
5.11 · UV, ozone/ORP, heating, cooling
| Equipment | Integration | Key fail-safe mechanism |
|---|---|---|
| UV steriliser | Pump + lamp on an outlet (no dimmable UV exists) | Flow interlock: UV powered only while the pump runs; a lamp-replacement reminder |
| Ozone + ORP | ORP probe → ozonizer outlet (~300–450 mV) | An ORP setpoint and a hard max-on timer; a dryer + a carbon scrubber |
| Heater | An outlet by temperature | An independent Inkbird as primary + the controller as over-temp protection; two small heaters |
| Chiller | An outlet by temperature | Its own thermostat + a wide hysteresis, a minimum cycle time |
| Fans | On-off / 0–10V by temperature | GHL — native dimming; Apex — VDM; an ATO is mandatory with evaporative cooling |
5.12 · Meters and monitors (the Milwaukee class)
Three classes: a spot tester (a one-off measurement), a monitor (probe + display, switches nothing), and a controller (monitor + relay). Only a "controller" can be a hardware cut-off.
| Device | Parameter | Class | Integration | Use |
|---|---|---|---|---|
| Milwaukee MC122 ★ | pH | Controller (1 relay) | Standalone; "seen" via outlet current monitoring | Reactor CO₂ cut-off; an independent kalk cut-off (~8.4) |
| Milwaukee MC125 ★ | pH + ORP | Controller (2 relays) | Standalone | CO₂ + ozone |
| American Marine Pinpoint ★ | pH / ORP / salinity / Ca | Monitor / controller | Standalone | An independent cross-check against the controller's probes |
| Hanna BL931700 ★ | pH (continuous) | Controller (relay + 4–20 mA) | The analogue output can feed a controller input | Industrial-grade CO₂ cut-off |
| Hanna Checkers | Alk / PO₄ / Ca | Spot tester | Standalone | Calibration and cross-checks |
| Seneye Reef ★ | NH₃, pH, temp, level, PAR | Monitor | HA via pyseneye (USB, local) | Free ammonia + PAR; the only monitor with clean local HA |
The independent cut-off principle (★). A standalone controller (Milwaukee, Pinpoint) with its own probe, power and logic will keep switching power even if the main controller hangs or goes offline. A software rule on the same Apex does not protect against a failure of the Apex itself. R. Holmes-Farley's recommendation: an MC122 as a kalk cut-off at pH 8.4. A single pH probe cannot be shared between two controllers — each needs its own. The main controller "sees" a standalone device trip via per-channel power monitoring.
5.13 · Leaks, sensors, feeders, wave scheduling
| Equipment | Apex | GHL | Hydros | HA |
|---|---|---|---|---|
| Leak detection | FMM + Leak Kit, interlock the return/ATO/CO₂ | Input (conductive only, not RO/DI) | Point/rope, up to 50 in a chain | Zigbee → outlets |
| Floats / optical / flow | FMM (4 ports) | Level/flow ports | Sense ports | ESP / Shelly |
| Feeders | Feed mode only pauses the pumps, it does not dispense food. The Eheim is a standalone timer+battery. A true "pause→feed→resume" cycle — only the Neptune AFS (AquaBus). | |||
| Wave scheduling | MXM / VDM | Tunze 1–10V; not Mobius-Vortech | WaveEngine (the broadest) | On-off |
6The closed-loop alkalinity circuit — the critical node
The most common cause of system deaths; fail-safes are mandatory
| # | Device | Strengths | Weaknesses | HA |
|---|---|---|---|---|
| 1 | Focustronic Alkatronic (or Alkatronic Pro) | the best long-term reputation, cheap reagent, standalone; the Pro adds pH+temp, heater/chiller and kalk control (but drops the BNC output) | HA is cloud scrape only; auto-mode without a dose limiter is dangerous | cloud |
| 2 | GHL KH Director | lab-grade precision, local operation, stops at the setpoint | finicky pH-electrode calibration, occupies 3 doser heads | DIY |
| 3 | Neptune Trident | the best local HA integration, falls back to baseline on error | the most failure reports, requires Apex, a short reagent shelf life | local |
| 4 | Hydros X10 + iV | the best fail-safe dosing settings, all-in-one | lengthy setup, probe drift | cloud |
| 5 | Kamoer KH Carer / D-D KH Manager | reagent is made yourself, a bridge into Hydros | cloud; a ~2.5 dKH error risk on a reagent batch change | cloud |
A hard daily-dose cap · a small reservoir · siphon-break protection · a high-pH cut-off (Milwaukee) · frequent small doses into a flow zone · regular manual cross-checks (Hanna/Salifert) · early reagent replacement · connect the Alk loop last.
Incidents: an Alkatronic in auto-mode raised the dose to 10,554 ml/day over a 0.25 dKH drop; a Trident reading "0" dumped ~4 L overnight; a Jebao DP-4 emptied the reservoir. Systems die from an unlimited dose triggered by a single erroneous measurement.
7Best in each category
By the totality of market data
| Category | Leader | Alternative | Rationale |
|---|---|---|---|
| Controller (reliability + locality) | GHL ProfiLux 4.1e | Neptune Apex A3 | the best hardware + a local web UI without the cloud |
| Controller (all-in-one Alk+dosing) | Hydros X10 + iV | — | controller + 4 dosers + auto-Alk in one case |
| Controller (openness / DIY) | AquaPi / ESPHome | reef-pi | native local HA, survives an HA outage |
| Closed-loop auto-Alk | Focustronic Alkatronic | GHL KH Director | years of reliability, cheap reagent, standalone |
| Doser (reliability) | GHL Doser 2.2 | Kamoer X1 Pro / Neptune DOS | 5+ years, repairable, a stepper drive (the Doser 3 has documented teething issues) |
| Doser (HA panel) | Red Sea ReefDose 4 | — | the best local read+write integration |
| Flow pump | Tunze Turbelle | EcoTech Vortech MP40 | the longevity benchmark, open 0–10V, no cloud |
| Auto top-off (reliability) | Tunze Osmolator 3 | XP Aqua Duetto2 | optical + thermal protection (no float), auto-recovery |
| Independent pH/ORP cut-off | Milwaukee MC122 | American Marine Pinpoint | a cheap standalone cut-off, independent of the controller |
| Controlled outlets | Hydros XP8 | Shelly Gen4 (for HA) | standalone, passive cooling, an alarm |
| Tablet dashboard | Home Assistant | Apex Fusion / local web | brings every brand into one local panel |
8Recommended builds
Three scenarios for different priorities; in all of them the Alk loop is protected by the mandatory mechanisms
Scenario 1 — "Maximum reliability and locality" (the baseline recommendation)
GHL ProfiLux 4.1e (or an Apex A3, if the Fusion dashboard is the priority) → closed-loop Alk on a standalone Focustronic Alkatronic + Dosetronic → Tunze flow (0–10V) or Vortech via MXM → a Tunze Osmolator 3 top-off + a mechanical valve → Shelly / EnergyBar outlets → an independent Milwaukee MC122 kalk cut-off → aggregation in Home Assistant on a wall-mounted tablet.
Scenario 2 — "An integrated single-case solution"
Hydros X10 + iV (controller + 4 dosers + closed-loop Alk) + WaveEngine for mixed-brand flow + an XP8 + an independent Milwaukee cut-off. Accept the lengthy iV setup and probe drift; HA is monitoring.
Scenario 3 — "Full DIY control"
AquaPi / ESPHome (a sensor and relay controller, native local HA) + a standalone Alkatronic + Shelly outlets + hardware fail-safes (Inkbird, floats, Milwaukee) + a custom panel on the HA WebSocket API.
Sources 2023–2026: reef2reef, ultimatereef, the CoralVue forum, YouTube (BRStv, Reef Builders), GitHub, manufacturer documentation, retail catalogues. Forum data is skewed toward complaints — this is the direction of consensus, not failure statistics. The "pro / con" quotes are representative and partly paraphrased. Prices are indicative. Claims were re-verified against primary sources; time-sensitive data (prices, ship dates, stock) was last checked June 2026.
An analytical report compiled from a multi-wave investigation · June 2026
Primary sources (authoritative references re-verified in the June 2026 wave):
- Neptune Apex A3 Series page
- Neptune COR return pump page
- Neptune DOS QuietDrive page
- GHL ION Director product page
- GHL KH Director product page
- GHL Doser 3 product page
- CoralVue WaveEngine pump compatibility chart
- Elwinmage/ha-reefbeat-component (GitHub)
- Red Sea ReefControl page
- Red Sea ReefATO+ page
- Tunze Osmolator 3 (3154) product page
- reef-pi (GitHub)
- roygabriel/hydrabridge-esp32 (GitHub)
- Focustronic Tracetronic product page
- Reefkeeping News — Reef Factory bankruptcy (May 2025)