Choosing ESP32 hardware and Sendspin or Squeezelite for Navidrome multi-room audio

Goal

I want to play music from my Navidrome server through several existing active speakers. Each speaker system would have an ESP32-based Wi-Fi audio endpoint connected to its 3.5 mm AUX input.

Home Assistant should provide playback and volume control. Synchronized multi-room playback is preferred, and I would eventually like to add Home Assistant voice control and announcements.

Current environment

  • Home Assistant version: 2026.7.4
  • Installation type: Container
  • Music Assistant version: Not installed yet
  • Navidrome version: 0.52.5
  • Network: HA server on Ethernet and ESP32 devices on 2.4 GHz wireless
  • Speakers: Various active speakers/soundbars

This is currently a planning question, so I don’t yet have failing YAML or logs to share.

Design under consideration

Navidrome
    │ OpenSubsonic
    ▼
Music Assistant
    │ Wi-Fi
    ▼
ESP32 audio endpoint
    │ stereo line-level output
    ▼
Active speaker AUX input

I have several ESP32-WROOM-32D boards, but they lack PSRAM, and I would need to buy separate DACs. Instead, I’m considering an integrated board such as the Sonocotta HiFi-ESP32 Plus, probably the ESP32-S3 version.

My leading firmware option is ESPHome with Sendspin. Squeezelite-ESP32 or Snapcast could be alternatives if they are more dependable.

Relevant documentation and projects I’ve found:

Sendspin appears closely integrated with Music Assistant and supports synchronization and delay adjustment, but it is still labeled a technical preview.

Questions

For anyone running ESP32 audio endpoints with Music Assistant:

  1. Is ESPHome with Sendspin reliable enough for everyday, long-running music playback?
  2. Would you choose the HiFi-ESP32 Plus S3, its classic ESP32 version, or another integrated board with stereo line output?
  3. Is Squeezelite-ESP32 currently a more dependable choice if synchronized playback is important?
  4. Have active speakers introduced enough internal latency to require per-player delay adjustment?
  5. Are Home Assistant announcements and returning to music afterward working reliably?

I plan to build and test one endpoint for my living room before purchasing several. I’d particularly appreciate first-hand experience with these hardware and firmware combinations.

1 yes been using it on various boards since the very early test versions. Now have multi room running most days.
2 I would build my own setups using esp32 s3. Look here for ideas, this will give you something very usable for little cash.

3 tried squeezelite a few times with ok results but esphome is better.
4 no idea as i don’t use them.
5 yes

You will need music assistant to handle the sendspin server. There are alternatives but I have little knowledge of them.

I have voice assistants with sendspin that perform ok. If you are not the sort for DIY the Vpe is probably the best place to start.