AI Engine Settings
The AI Engine tab is the brain of your dIKta.me experience. It controls which Speech-to-Text (STT) and Large Language Model (LLM) engines handle each pipeline. The tab has a master-detail layout — select a category on the left to configure it on the right.
TIP
**Cloud vs. Local**: You can switch the active environment directly from the Control Panel overlay without opening settings.
**Cloud vs. Local**: You can switch the active environment directly from the Control Panel overlay without opening settings.
Environment Routing
- Cloud Route (Default): Uses third-party API providers over the internet.
- STT Provider: Deepgram (streaming) or OpenAI Whisper (batch), depending on which keys you have configured.
- LLM Provider: Choose from Gemini, Anthropic (Claude), OpenAI (GPT), OpenRouter, or Requesty — whichever keys you have saved.
- Local Route (On-Device): Uses offline AI modules. Your audio never leaves your machine.
- STT Provider: Whisper.net — downloads an ONNX model the first time it runs.
- LLM Provider: Ollama — communicates with local models running on your hardware.
Model Selection
Once an environment is active, use the model dropdowns to pick the exact model for each pipeline. For example, with Gemini selected you can choose between gemini-2.5-flash (fast, default) or a more capable model variant.
Changing the model dropdown affects every pipeline execution: speed, accuracy, and cost all shift accordingly.
Sub-sections
The AI Engine settings page contains the following sub-sections:
| Section | What it configures |
|---|---|
| API Keys | Save and manage API keys for all cloud providers |
| Speech to Text | Active STT provider and model for Cloud and Local routes |
| Language Model | Active LLM provider and model for Cloud and Local routes |
| Text to Speech | Active TTS provider and voice settings |
| Chat | Quick Chat system prompt and history settings |
| System Monitor | Ollama health, GPU status, and model cache |