Home Assistant Voice Assistant — Complete Guide

In this tutorial, you'll learn about Home Assistant Voice Assistant. We cover key concepts, practical examples, and best practices.

Hook

You configure the Home Assistant Voice Assistant (Assist) with a microphone and speaker. You say "Turn on the kitchen light." Assist responds "I'm sorry, I couldn't process that" or just displays "Listening..." forever. The voice pipeline is not completing.

The Wrong Way

Switching between multiple voice pipelines (Whisper, Cloud, Piper) without testing each component individually does not identify which part fails.

# BAD: Changing pipeline without testing
# Switched from Cloud to Whisper to Piper in 5 minutes
# No component was tested individually

Pipeline tested: Cloud → Cloud, Whisper → Cloud, Piper → Cloud
Result: All fail at different stages

You need to isolate the failing component.

The Right Way

Test each stage of the voice pipeline independently.

# 1. Test speech-to-text (STT)
curl -X POST -H "Authorization: Bearer <token>" \
  -F "file=@test.wav" \
  -F "language=en" \
  http://hass.local:8123/api/stt/cloud

{"text": "turn on the kitchen light"}

If STT works, test text-to-speech (TTS):

# 2. Test TTS
curl -X POST -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{"platform": "cloud", "message": "Turning on the light", "media_player_entity_id": "media_player.kitchen_speaker"}' \
  http://hass.local:8123/api/services/tts/cloud_say

{"message": "TTS generated and played"}

# 3. Test Assist intent handling
curl -X POST -H "Authorization: Bearer <token>" \
  -H "Content-Type: application/json" \
  -d '{"name": "Turn on the kitchen light"}' \
  http://hass.local:8123/api/services/conversation/process

{
  "response": {
    "speech": {
      "plain": {
        "speech": "Turned on kitchen light"
      }
    }
  }
}

# 4. If pipeline assembly fails, check the pipeline config:
# Settings → Voice Assistants → Assist → Pipeline → Select correct STT/TTS/Conversation agents

Prevention

• Test STT, TTS, and conversation processing individually before combining.
• Use the Cloud pipeline for initial setup (it works out of the box).
• For local assistants, install Wyoming protocol add-ons (Whisper, Piper) and assign them in the pipeline.
• Keep microphone and speaker volumes at reasonable levels.
• Use a push-to-talk button for initial testing before enabling wake-word detection.

Advanced Troubleshooting

Check the Logs

Most TOOL errors are logged to stdout or a dedicated log file. Check your logs first:

# Check system logs
journalctl -u tool --since "1 hour ago"

# Or check the application log
tail -50 ~/.tool/logs/error.log

Test with a Minimal Example

Create the simplest possible tool configuration to verify the base setup works:

tool --version
tool --help

If the minimal test passes, add configuration options one at a time until you find the breaking change.

Common Configuration Mistakes

• Using the wrong file path or URL in configuration
• Forgetting to restart TOOL after changing config files
• Mixing tabs and spaces in YAML configuration files
• Setting incorrect permissions on configuration directories

When to Reinstall

If none of the above resolves the issue, consider a clean reinstall:

# Backup your configuration
cp -r ~/.tool ~/.tool.bak

# Remove and reinstall
# Follow the official TOOL installation guide

This ensures you start from a known good state and can isolate the issue.

Common Mistakes with assistant voice

1. Misunderstanding that String is [Char] with poor performance for large text operations
2. Using foldl instead of foldl' causing stack overflow on large lists
3. Forgetting deriving (Show, Eq) on custom data types needed for debugging

These mistakes appear frequently in real-world HOME code. DodaTech's contributors have identified these patterns through analysis of open-source projects and production systems.

Practice Exercise

Write a pure function that safely divides two integers using Maybe, then test it with edge cases like division by zero and negative numbers.

This exercise reinforces the concepts covered in this guide. Try implementing it before checking online solutions.

FAQ

DodaTech — a voice assistant that listens and responds.