Project: Offline Voice Command Node - BOM, firmware, build notes

[polygoku]

Offline Voice Command Node — MAKHEX build log

Current status: Blueprint, not an electrically or mechanically qualified kit.

Purpose: Experiment with local push-to-talk and spoken status responses.

Candidate bill of materials

• 1x /esp32-s3-devkit — Processor
• 1x /i2s-mems-mic — Audio input
• 1x /i2s-amplifier — Speaker driver
• 1x /usb-c-power-kit — Stable supply

Proposed build milestones

1. Verify microphone and amplifier independently.
2. Check memory and firmware SDK compatibility.
3. Build push-to-talk with a visible recording indicator.
4. Benchmark a small on-device command recognizer.
5. Add error recovery and timeout handling.

Improve this guide

• Suggest validated supplier or component substitutions
• Share board revision and software environment
• Post reproducible firmware builds, wiring diagrams and measurements
• Document errors, successful tests and unresolved risks

Engineering caution: General-purpose offline speech recognition is not practical on all ESP32 modules; verify model memory needs.

Please be specific and label unverified approaches.