🔧 Transformer Plugins (Runtime Modifiers)

Transformer plugins operate independently of personas and provide fine-grained control over OVOS’s internal processing pipeline. They are not part of the persona framework but can synergize with it.

Key Details:

• Scope: Transformers apply within the OVOS core pipeline—not inside personas (though solver plugins can use them internally if desired).
• Independence: Transformers and personas are separate systems. However, future enhancements may allow dynamic interaction between the two (e.g., a persona adjusting transformer settings).

Key Integration Points:

• Utterance Transformers: Operate between STT (Speech-to-Text) and NLP (Natural Language Processing).
• Dialog Transformers: Operate between NLP and TTS (Text-to-Speech).

---

Examples of Transformer Plugins Using AI Agents

✅ OVOS Transcription Validator

This plugin validates the output of STT engines using a language model to filter out incorrect or incoherent transcriptions before they are passed to NLP.

How It Works:

1. Receives an STT transcript and its language code.
2. Sends both to an LLM prompt (local or via Ollama).
3. Gets a True or False response based on utterance validity.

Configuration Snippet (mycroft.conf):

"utterance_transformers": {
  "ovos-transcription-validator-plugin": {
    "model": "gemma3:1b",
    "ollama_url": "http://192.168.1.200:11434",
    "prompt_template": "/path/to/template.txt",
    "error_sound": true,
    "mode": "reprompt"
  }
}

Use Case: Prevent skills from being triggered by invalid STT output like "Potato stop green light now yes."

---

🗣️ Dialog Transformer

This plugin rewrites assistant responses based on a persona-style prompt, enabling tone or complexity adjustments.

Example Prompt Use Cases:

• "Rewrite the text as if you were explaining it to a 5-year-old"
• "Explain it like you're teaching a child"
• "Make it sound like an angry old man"
• "Add more 'dude'ness to it"

Configuration Snippet (mycroft.conf):

"dialog_transformers": {
  "ovos-dialog-transformer-openai-plugin": {
    "rewrite_prompt": "rewrite the text as if you were explaining it to a 5-year-old"
  }
}

This plugin often leverages LLMs through solver plugins but operates after the main dialog logic, adjusting the final output.