Capability note 02
Meta Muse Gadgets voice support, without the guesswork.
Voice input, text replies and spoken output are three different capabilities.
Source-checked 2026-10-05 No hardware testing
A text reply with no sound can be expected.
The ESP32 documentation describes text replies by default. Spoken replies require a separate text-to-speech path. A microphone or speaker on the board does not establish working, out-of-the-box two-way voice.
Read the pinned official voice guidance ↗Check each part of the journey.
Voice input
The board and firmware capture speech, usually through a push-to-talk action.
Text reply
The request reaches Muse and a supported display shows the response.
Spoken output
An additional TTS integration turns response text into audio on suitable hardware.
What is the last stage that worked?
Start with the output capability.
Check whether your exact firmware includes a TTS integration. Stock text replies do not become speech simply because the board has a speaker. Some constrained boards cannot use the same audio path.
See spoken-output conditions →Check the build and the failure stage.
PR #70 addresses queue allocation in a Core2 contributor’s task-startup report. A generic “start failed” message can have several causes; it does not by itself diagnose that exact problem. We have not reproduced the fix.
Read the merged allocation change ↗Keep the later response stage separate.
Issue #87 describes this on Waveshare ESP32-S3-Touch-AMOLED-1.75C with a reported 031f65b build. The existence of PR #70 does not establish that this later-stage problem is fixed. Keep the precise board, firmware revision and last successful stage together.
Read the specific report and current discussion ↗This is a source guide, not a device diagnosis.
We do not inspect your device, account or logs. Do not share SDK Tokens, Wi-Fi credentials, private recordings or full personal logs. Use the official setup and troubleshooting paths for your exact board.