[shelly] Add support for Duo E27 G3 and Multicolor Bulb E27 G3 by markus7017 · Pull Request #20909 · openhab/openhab-addons

markus7017 · 2026-06-07T20:26:09Z

Description

[New Device Support]

Add support for the Shelly Duo Bulb E27 Gen3 (shellyplusduobulb) and Shelly Multicolor Bulb E27 Gen3 (shellypluscolorbulb). Both devices use the Gen2 RPC API. While integrating them, the shared light-handling code used by all Shelly light devices was reviewed and adjusted so each bulb's actual color profile (CCT-only, RGB+CCT, or RGBW2's fixed 4-channel color+white) is handled correctly — this also affects the existing Gen1 Duo, Duo RGBW and Vintage bulbs, which share the same code paths. README documentation was updated for both new thing types.

Features:

Add support for the Shelly Duo Bulb E27 Gen3: tunable-white (CCT) bulb with brightness, color temperature and auto-on/off timer control
Add support for the Shelly Multicolor Bulb E27 Gen3: RGB+CCT bulb whose color and white channel groups are both always active and can be controlled independently at the same time

Fixes:

Corrected several control calls (status fetch, brightness, turn, auto-on/off timer) that used the wrong RPC method for the Multicolor Bulb G3, so they did not work
Removed a false "with Power Meter" claim from both new thing-type descriptions
Reviewed the light status/command handling shared by all bulb and RGBW2 devices (Gen1 and Gen2) and adjusted it so each device's actual color profile is respected, instead of assuming the RGBW2 shape everywhere — this also benefits the existing Gen1 Duo, Duo RGBW and Vintage bulbs

Changes:

Each bulb model's color profile (CCT-only Duo Bulb G3, RGB+CCT Multicolor Bulb G3, RGBW2's fixed 4-channel color+white) is now handled on its own terms

Testing

How to verify on hardware

Channel state updates arrive at the normal polling interval, not instantly on change; this matches the existing Duo/RGBW2 bulb behaviour on this branch.
No meter channels are created for either new device; their status payload carries power/energy fields, but the binding does not yet wire them into channels.
Duo Bulb G3: verify brightness, color temperature and auto-on/off timer control.
Multicolor Bulb G3: verify RGB/HSB control in color mode, brightness/color-temperature control in white mode, and that switching between the two modes updates the right channel group.
Existing Gen1 Duo, Duo RGBW and Vintage bulbs: verify behavior is unchanged from before this PR (power/brightness/color-temperature channels and color/white mode switching still work as before).

Acceptance criteria

Implementation done
Documentation is up-to-date
Verified by Community

Closing

[shelly] Support for Duo bulb Gen3 #20851

Backport assessment

New device support bundled with a review/fix pass on the shared light-handling code it depends on; not backport-eligible.

lsiepel

Thanks, LGTM

This is still marked WIP, how to proceed. We only have about a week left for code freeze. There is also a conflict to solve.

markus7017 · 2026-06-12T21:52:20Z

@lsiepel Yes, I already spend some time, but couldn't get it stable - fixing here breaks there. This will not go into 5.2. I'm focusing of bug fixes and "easy integrations" of new devices (there are 15 more...). WS90 and LORA might be good candidates because almost done.

openhab-bot · 2026-07-11T08:13:07Z

This pull request has been mentioned on openHAB Community. There might be relevant details there:

https://community.openhab.org/t/shelly-binding/56862/5689

openhab-bot · 2026-08-03T11:16:55Z

This pull request has been mentioned on openHAB Community. There might be relevant details there:

https://community.openhab.org/t/shelly-beta-with-plus-pro-mini-and-blu-support/139554/1

matmai · 2026-08-04T10:22:57Z

I am also willing to test if you could provide a jar.

andrewfg

Just a few comments concerning terminology

lsiepel · 2026-08-09T19:29:32Z

@markus7017 there is an offer to get this tested by @matmai, couldyou provide a jar where the conflicts and feedback have been resolved ?

Copilot

Pull request overview

Adds Gen3 Shelly Duo and Multicolor bulbs using the Gen2 RPC API.

Changes:

Registers both bulb types and channel definitions.
Adds CCT/RGB DTO and RPC handling.
Adds profile and deserialization tests.

Reviewed changes

Copilot reviewed 14 out of 14 changed files in this pull request and generated 13 comments.

Show a summary per file

File	Description
`ShellyApiConfigurationTest.java`	Improves setup failure reporting.
`Shelly2DuoBulbProfileTest.java`	Tests bulb DTO deserialization.
`ShellyDeviceProfileTest.java`	Tests new profile flags.
`shellyGen2_lights.xml`	Defines new thing types.
`shellyGen1_lights.xml`	Adds Duo temperature range type.
`shelly.properties`	Adds labels and descriptions.
`ShellyHandlerFactory.java`	Removes an unused field.
`ShellyDevices.java`	Registers device identifiers and groups.
`ShellyChannelDefinitions.java`	Creates CCT channels dynamically.
`ShellyLightHandler.java`	Handles Gen3 bulb state and commands.
`Shelly2ApiRpc.java`	Adds bulb RPC dispatch.
`Shelly2ApiJsonDTO.java`	Adds RGB/CCT RPC DTOs.
`Shelly2ApiClient.java`	Maps bulb configuration and status.
`ShellyDeviceProfile.java`	Adds Gen3 bulb capability flags.

Suppressed comments (2)

bundles/org.openhab.binding.shelly/src/main/resources/OH-INF/thing/shellyGen2_lights.xml:26

The description still claims a power meter, contradicting the PR description and the implementation, which does not map the bulb's power/energy fields to meter channels. Remove the claim so users are not promised unavailable channels.

		<description>Shelly Multicolor Bulb E27 - Dimmable RGB+CCT Bulb with Power Meter</description>

bundles/org.openhab.binding.shelly/src/main/resources/OH-INF/i18n/shelly.properties:84

The localized description reintroduces the false power-meter claim that the PR description says was removed. This string will override the corrected thing XML in localized UIs.

thing-type.shelly.shellypluscolorbulb.description = Shelly Multicolor Bulb E27 Gen3 (Dimmable RGB+CCT Bulb with Power Meter)

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

markus7017 · 2026-08-09T21:50:13Z

@lsiepel

@markus7017 there is an offer to get this tested by @matmai, couldyou provide a jar where the conflicts and feedback have been resolved ?

That's to early. We discussed

bug fix pr
feature pr
oh core light
duo g3

lsiepel · 2026-08-09T22:05:19Z

@lsiepel

@markus7017 there is an offer to get this tested by @matmai, couldyou provide a jar where the conflicts and feedback have been resolved ?

That's to early. We discussed

bug fix pr

feature pr

oh core light

duo g3

According to the core light pr startpost it depends on this PR.
According to the feature pr start post it also depends on this. So it looks like this needs to go first.

so after the almost finished bug fix I assumed this PR is next. If not please update the above linked PR start post. And verify with @andrewfg about the core pr dependency

andrewfg · 2026-08-09T22:46:50Z

My PR will only come after @markus7017 three light related PRs (including this one) are finished.

lsiepel · 2026-08-23T21:15:51Z

Be carefull when rebasing, when done, i'll run another review round. Are there any device owners who mentioned they are willing to test?

markus7017 · 2026-08-25T13:52:15Z

The commit history has been rebased and reorganized into review-friendly, cherry-pickable commits (one per fix/feature/test group).

This PR adds support for the Shelly Duo Bulb E27 Gen3 and Shelly Multicolor Bulb E27 Gen3. While integrating them, the shared light-handling code used by all Shelly light devices was reviewed and fixed — this also covers the existing Gen1 Duo, Duo RGBW and Vintage bulbs, since they share the same code paths. Each bulb model's color profile (CCT-only, RGB+CCT, or RGBW2's fixed color+white) is now handled on its own terms instead of assuming one shape for all of them.

A build is available for testing:

Jar: org.openhab.binding.shelly-5.3.0-SNAPSHOT-duog3.jar
README: README.md
Install instructions: READMEbeta

@matmai @andrewfg would appreciate your help testing this on real hardware — Duo Bulb G3, Multicolor Bulb G3, and, if you have one, an existing Gen1 Duo/Duo RGBW/Vintage bulb to confirm nothing regressed there.

markus7017 · 2026-08-25T14:06:31Z

@lsiepel Please trigger an AI review

markus7017 · 2026-08-25T15:47:17Z

This gives an overview of the different color profiles, device types and how they map to channels
light-devices-overview.md

lsiepel

Here are the review results.

This review was AI-assisted.

mstormi · 2026-08-25T19:50:53Z

Did some testing, here's the feedback (if I understood correctly you want it here).

FWIW, I didn't use discovery but statically created a bulb thing.

So: can control brightness but not color temperature. The latter channel has no effect.

And the brightness channel seems to reset itself to 0 or 100 at times. Not sure if my testing method or some other bug that still references the old (standard binding) channel is still active (cannot restart my system).

markus7017 · 2026-08-25T23:04:14Z

@matmai Please use updated build and try to use discovery initially, maybe the are issues when using the .things file. Delete the thing (also from jsondb to avoid some old stuff to get restored)

andrewfg · 2026-08-27T18:51:29Z

@markus7017 I could not really tell from your code, but it looks like you are reclassifying the RGBWWPM rgbcct mode to be one "light" (essentially the same as a Gen 3 full color bulb) rather than a separate rgb primary with a cct secondary. Is that correct?

markus7017 · 2026-08-27T22:09:54Z

@andrewfg
Correct, and intentional — they're different hardware shapes that happen to share the "rgbcct" name. Pro RGBWW PM's rgbcct firmware profile is two independent components (RGB primary + CCT secondary). The Gen3 Multicolor
Bulb's rgbcct:0 is a single RPC component per Shelly's own Gen3 docs - no separate RGB/CCT components
exist on that device, so modeling it as one light is the accurate representation, not a simplification.

Maybe to reduce future confusion (not implemented yes): the isRGBCCT flag name is what's colliding with the Pro RGBWW PM's rgbcct profile string in reviewers' heads even though the code paths never overlap (isRGBCCT is gated behind isDuo everywhere). If you want to preempt this coming up again, renaming the flag to something bulb-specific (e.g. isMulticolorBulb) or adding a one-line Javadoc contrast note on the field would remove the ambiguity without touching behavior.

lsiepel

This is probably the last one

lsiepel · 2026-08-28T12:41:33Z

+            Shelly2RpcRequestParams params = new Shelly2RpcRequestParams();
+            params.id = 0;
+            params.mode = SHELLY_MODE_COLOR.equals(mode) ? SHELLY_RGBCCT_MODE_RGB : SHELLY_RGBCCT_MODE_CCT;
+            apiRequest(SHELLYRPC_METHOD_RGBCCT_SET, params, String.class);


RGBCCT.Set requires at least one of on or brightness, so this mode-only request is invalid. The same precondition is missed by the setLightParms() RGBCCT branch when it sends only rgb or ct; the temperature path first sends this request and then a ct-only request, so switching to white/CCT can fail before the temperature is applied. Please ensure every RGBCCT.Set request includes on/brightness (or combine mode and value into one valid request), and make the transport-stub test reject payloads that violate the documented requirement.

Fixed in 9823a91 - the handler now folds the mode switch into the same RGBCCT.Set request as the new rgb/ct value (one request instead of mode-only + value-only), and the RPC layer adds on=true whenever a request would otherwise carry neither on nor brightness. setLightMode() goes through the same path. The transport-stub test now rejects any RGBCCT.Set payload without on/brightness, and new tests cover the combined mode+ct / mode+rgb requests. (CCT.Set is unaffected: per the docs it accepts ct alone.)

markus7017 · 2026-08-29T09:42:20Z

@lsiepel changes applied, ready for the next round:

Fix: switching the Multicolor Bulb G3 between color and white mode and setting the color/temperature now goes out as a single valid RGBCCT.Set request; every RGBCCT.Set carries on/brightness as required by the API (9823a91)
Fix: the Gen3 bulbs' brightness/power state no longer drops to OFF / 0 % on every status poll while the bulb is on (d8d24f0) - @mstormi this is very likely the "brightness resets itself to 0 or 100" you observed, please re-test with the build below
Cleanup from a final self-review, no functional change (a3a4b4d)

Branch is rebased on current main, full build incl. tests is green.

Updated build: org.openhab.binding.shelly-5.3.0-SNAPSHOT-duog3.jar

mstormi · 2026-08-29T11:13:30Z

this is very likely the "brightness resets itself to 0 or 100" you observed, please re-test with the build below

Yes that seems to have fixed it.
The other problem (color temp not changeable) persists, though.

…profile flags Signed-off-by: Markus Michels <markus7017@gmail.com>

Signed-off-by: Markus Michels <markus7017@gmail.com>

…d mode switching Signed-off-by: Markus Michels <markus7017@gmail.com>

…nitions Signed-off-by: Markus Michels <markus7017@gmail.com>

…tch handling Signed-off-by: Markus Michels <markus7017@gmail.com>

Signed-off-by: Markus Michels <markus7017@gmail.com>

- derive isRGBCCT from the thing type at init time (Multicolor Bulb is always rgbcct:0), runtime config detection stays as confirmation - drop the Gen1-Bulb-only mode channel update for the Multicolor Bulb (it has no mode channel) - remove redundant isDuo guard and no-op list write in Shelly2ApiClient - rename Shelly2DuoBulbProfileTest to Shelly2DuoBulbDtoTest (it tests DTO mapping only), use the Multicolor thing type in color-mode tests, drop comments/FQNs in unit tests - README: one sentence per line for the Multicolor Bulb paragraph; trim comments Signed-off-by: Markus Michels <markus7017@gmail.com>

…tus blanking, CCT.Set guard - updateRGBW: skip partial status notifications without rgb instead of dereferencing null - applyBulbLightSettings: preserve the persisted light status across profile refreshes - CCT.Set: don't send an empty request (device requires on, brightness or ct) - RGBCCT.Set: repeat the current power state instead of forcing on when only rgb/ct/mode is sent - Multicolor Bulb: white#temperature reports UNDEF in color mode; Gen3 bulbs push Kelvin - ShellyColorUtils copy constructor now carries the mode Signed-off-by: Markus Michels <markus7017@gmail.com>

…eep Gen1 bulb behaviour - color#hsb/red/green/blue now switch the Multicolor Bulb to color mode, white#temperature switches back - Gen3 bulbs have no power channel: brightness 0 turns the light off, color picker brightness is applied - Gen1 Bulb/Duo/Vintage/Duo RGBW keep their power channel and brightness semantics Signed-off-by: Markus Michels <markus7017@gmail.com>

…r Gen3 bulbs - Gen3 Duo/Multicolor Bulb white#temperature is Number:Temperature 2700..6500 K - Gen1 devices keep the Dimmer based whiteTemp channel type - Thing descriptions aligned with i18n properties, Duo Bulb category Lightbulb - README documents the channel type and mode switching Signed-off-by: Markus Michels <markus7017@gmail.com>

…ests - Shelly2RGBCCTStatus nested in Shelly2DeviceStatusResult like the other status DTOs - SHELLYDT_COLORBLBG3 renamed to SHELLYDT_COLORBULBG3 (value unchanged) - Discovery tests cover the Gen3 bulb service names and model IDs Signed-off-by: Markus Michels <markus7017@gmail.com>

markus7017 · 2026-08-29T13:47:41Z

Pushed a self-review round for both Gen3 bulbs (f3a72e23b0). Gen1 Bulb/Duo/Vintage/Duo RGBW behaviour is unchanged, all Gen3-specific handling is gated on the Gen3 bulb thing types.

Fixes

Fix: A partial status push from the Multicolor Bulb (e.g. only brightness) no longer causes an exception; the channels are updated from the values that were sent.
Fix: color#hsb, color#red/green/blue now switch the Multicolor Bulb to color mode instead of being ignored while the bulb is in white mode; white#temperature switches back to white mode.
Fix: Changing only the color or color temperature no longer turns a switched-off bulb on; the current power state is kept.
Fix: Setting brightness 0 turns the Gen3 bulbs off in both color and white mode.
Fix: white#temperature no longer shows a stale value while the Multicolor Bulb is in color mode (reports UNDEF).
Fix: Color temperature commands without a value change are no longer sent as an empty request the device rejects.

Changes

Change: white#temperature for the Duo Bulb G3 and Multicolor Bulb G3 is now a Number:Temperature channel in Kelvin (2700..6500 K) instead of a percentage. Existing Gen1 bulbs keep the percent based channel.
Change: The Gen3 bulbs have no control#power channel; use white#brightness (0 = off) or color#hsb.
Change: Thing descriptions aligned with the i18n texts, Duo Bulb category is now Lightbulb.

README updated accordingly. Test build: org.openhab.binding.shelly-5.3.0-SNAPSHOT-duog3.jar in the usual myfiles location.

markus7017 · 2026-08-29T15:36:29Z

@mstormi please re-test

mstormi · 2026-08-30T16:38:57Z

yes color temp is working now, thanks

lsiepel

Thanks, LGTM

awaiting confirmation by markus

matmai · 2026-08-31T21:06:13Z

I also have also tested a S3BL-C010007AEU bulb with org.openhab.binding.shelly-5.3.0-SNAPSHOT-duog3.jar (5.3.0.202608291343) on openHAB 5.3.0 Snapshot Build 5566.

It is discovered in the inbox
brightness works
color#HSB works
``clor#full` works

white#temperature works not fully:

Sometimes it needs multiple changes until the bulb changes its temperature. From the web UI it always works.
The values in Kelvin are not the same as in the web UI, examples:
- 2700K (OH) was set to 2993K (web)
- 5000 (OH) was set to 5273K (web)
- 6500L (web) was set to 6230K (OH)

BTW: SHBLB-1 was not supported by this bundle

markus7017 self-assigned this Jun 7, 2026

markus7017 added enhancement An enhancement or new feature for an existing add-on work in progress A PR that is not yet ready to be merged labels Jun 7, 2026

This was referenced Jun 7, 2026

[shelly] Add support for Shelly Pro RGBWW PM #19227

Merged

[shelly] Add Plus/Pro Dimmer models (Gen3/Gen4): 1PM, 2PM, 10V, DALI #19226

Open

lsiepel reviewed Jun 10, 2026

View reviewed changes

Comment thread ...b.binding.shelly/src/main/java/org/openhab/binding/shelly/internal/ShellyHandlerFactory.java

lsiepel reviewed Jun 11, 2026

View reviewed changes

markus7017 added the additional testing preferred The change works for the pull request author. A test from someone else is preferred though. label Jul 11, 2026

markus7017 force-pushed the shelly_duog3 branch from d202a83 to b6afa2e Compare July 19, 2026 21:57

markus7017 mentioned this pull request Jul 19, 2026

[shelly] Support for Duo bulb Gen3 #20851

Open

markus7017 force-pushed the shelly_duog3 branch from b6afa2e to d202a83 Compare July 20, 2026 04:38

markus7017 mentioned this pull request Jul 20, 2026

[shelly] PR review orchestration #20528

Open

markus7017 mentioned this pull request Aug 2, 2026

[shelly] Fix Plus RGBW PM profiles and zero brightness handling #21256

Merged

3 tasks

andrewfg reviewed Aug 5, 2026

View reviewed changes

andrewfg added awaiting other PR Depends on another PR and removed awaiting other PR Depends on another PR labels Aug 5, 2026

andrewfg mentioned this pull request Aug 5, 2026

[shelly] Implement OH Core LightModel #21286

Draft

13 tasks

lsiepel requested a balanced review from Copilot August 9, 2026 19:29

Copilot started reviewing on behalf of lsiepel August 9, 2026 19:30 View session

Copilot AI reviewed Aug 9, 2026

View reviewed changes

andrewfg reviewed Aug 17, 2026

View reviewed changes

Comment thread ...inding.shelly/src/main/java/org/openhab/binding/shelly/internal/api/ShellyDeviceProfile.java Outdated

lsiepel requested changes Aug 25, 2026

View reviewed changes

lsiepel requested changes Aug 28, 2026

View reviewed changes

markus7017 force-pushed the shelly_duog3 branch from 85a9c7b to a3a4b4d Compare August 29, 2026 09:41

markus7017 added 12 commits August 29, 2026 15:10

[shelly] Register Duo Bulb G3 and Multicolor Bulb G3 thing types and …

d3ab4e0

…profile flags Signed-off-by: Markus Michels <markus7017@gmail.com>

[shelly] Add RGBCCT DTO fields and RPC method constants

b08da8f

Signed-off-by: Markus Michels <markus7017@gmail.com>

[shelly] Add Gen2 RPC dispatch for cct:0/rgbcct:0 light components

a4461e9

Signed-off-by: Markus Michels <markus7017@gmail.com>

[shelly] Wire Duo/Multicolor Bulb G3 status parsing, timer fields, an…

708d541

…d mode switching Signed-off-by: Markus Michels <markus7017@gmail.com>

[shelly] Create Duo/Multicolor Bulb G3 channels and channel-type defi…

3110ae4

…nitions Signed-off-by: Markus Michels <markus7017@gmail.com>

[shelly] Route light commands/status through white-group and mode-swi…

7300df1

…tch handling Signed-off-by: Markus Michels <markus7017@gmail.com>

[shelly] Document Duo Bulb G3 and Multicolor Bulb G3 in README

aa24e0b

Signed-off-by: Markus Michels <markus7017@gmail.com>

markus7017 force-pushed the shelly_duog3 branch from a3a4b4d to f3a72e2 Compare August 29, 2026 13:45

lsiepel approved these changes Aug 30, 2026

View reviewed changes

Uh oh!

Uh oh!

Conversation

markus7017 commented Jun 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Testing

Acceptance criteria

Closing

Backport assessment

Uh oh!

Uh oh!

lsiepel left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

markus7017 commented Jun 12, 2026

Uh oh!

openhab-bot commented Jul 11, 2026

Uh oh!

openhab-bot commented Aug 3, 2026

Uh oh!

matmai commented Aug 4, 2026

Uh oh!

andrewfg left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

lsiepel commented Aug 9, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

markus7017 commented Aug 9, 2026

Uh oh!

lsiepel commented Aug 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

andrewfg commented Aug 9, 2026

Uh oh!

Uh oh!

lsiepel commented Aug 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

markus7017 commented Aug 25, 2026

Uh oh!

markus7017 commented Aug 25, 2026

Uh oh!

markus7017 commented Aug 25, 2026

Uh oh!

lsiepel left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mstormi commented Aug 25, 2026

Uh oh!

markus7017 commented Aug 25, 2026

Uh oh!

andrewfg commented Aug 27, 2026

Uh oh!

markus7017 commented Aug 27, 2026

Uh oh!

lsiepel left a comment

Choose a reason for hiding this comment

markus7017 commented Jun 7, 2026 •

edited

Loading

lsiepel left a comment •

edited

Loading

lsiepel commented Aug 9, 2026 •

edited

Loading

lsiepel commented Aug 23, 2026 •

edited

Loading

lsiepel left a comment •

edited

Loading