OpenClaw has simplified Apple Watch setup by folding voice into the normal connection flow. PR #139712, "fix(watch): include voice in normal setup," landed at 07:44 UTC on September 6, 2026.
Before this change, connecting an Apple Watch exposed two setup actions. The ordinary direct connection did not include voice, so users had to discover a separate "Enable Standalone Voice" action before they could talk to OpenClaw from the Watch.
The new flow replaces that split path with one Connect Apple Watch action that requests the existing voice-capable pairing profile.
What Changed
The iOS and Watch setup flow now uses a single voice-capable pairing profile. The alternate node-only selection and the includeVoice parameter were removed instead of being kept as a hidden default.
The PR also updates Watch guidance, setup documentation, and native string inventory together. That matters because setup text, native UI, and generated localization expectations can drift when a feature is changed in only one layer.
Security boundaries stay narrow. The authenticated iPhone still needs admin authority to issue a one-time setup code. The Watch receives its own node credential and exactly operator.read and operator.talk. Phone credentials and provider credentials are not copied.
There are no Gateway, provider, media transport, database schema, or protocol-version changes.
User Impact
The user-facing impact is simple: normal Watch setup now includes voice without a separate enable setting.
Users still need to press Start and grant microphone permission before calls and microphone capture work. Existing direct-connection off and forget controls remain available.
For older node-only Watch setups, the migration path is explicit rather than silent. Users can run Connect Apple Watch again to authorize voice. Existing grants are not automatically expanded, and revoked permissions are not automatically restored.
That is the right balance for a wearable voice feature. The setup gets easier, but permission changes still require visible user action.
Why It Matters
Apple Watch is one of the places where OpenClaw should feel immediate. If a user pairs the Watch and then still has to find a second voice enable action, the flow feels half-connected.
By making voice part of the normal connection profile, OpenClaw removes that extra discovery step. It also gives documentation and native strings the same mental model: Connect Apple Watch means connect the Watch for reading and talking.
This is a small product change, but a meaningful onboarding improvement for people who want OpenClaw available from their wrist.
Evidence From The PR
The PR reports changed-scope checks covering native protocol behavior, formatting, Swift lint, and the native-state schema guard. Source-native localization verification passed, and the exact-commit CI run passed on the reviewed source head.
The most useful validation is the UI evidence. The existing Settings-to-Watch UI test failed on the baseline because Connect Apple Watch was absent, then passed on the candidate. A final after-only rerun built the exact head and passed the same test, including opening Message Delivery.
The PR is careful about the boundary of that proof. It validates the Settings UI and setup flow in synthetic iPhone Simulator captures. It does not claim physical Watch pairing, microphone behavior, speaker behavior, or a new provider session.
Even with that limitation, the merged behavior is clear: the normal Apple Watch setup path now includes voice, with tighter UX and the same explicit credential boundaries.