AI Providers Configuration
Connect cloud APIs or local instances to power the prompt generator and design coach
AI Level Scatter doesn’t lock you into a single proprietary AI model. It features a modular AI API Client Layer that allows you to connect to a variety of artificial intelligence models, including cloud-based LLMs (like Google Gemini, OpenAI GPT-4o, and Anthropic Claude) and local offline instances (using Ollama).
Configure and manage your API keys, base endpoints, model names, and connection tests in one unified interface.
AI Settings Manager
Open the configuration window via Tools → Decnet → AI Settings Manager.
From this panel, you can choose which API provider is active, toggle offline modes, input credentials, and test connection latency:
┌──────────────────────────────────────────────┐
│ AI Settings Manager │
├──────────────────────────────────────────────┤
│ Active Provider: [ Google Gemini ] │
│ Model Name: [ gemini-1.5-pro-latest ] │
│ API Key: [ ******************** ] │
│ │
│ [ Test Connection ] Latency: 120ms (OK) │
└──────────────────────────────────────────────┘
Provider Documentation
Select a sub-page below to configure your specific AI provider:
- Google Gemini Setup: Standard provider for text-to-preset compilation and design coach queries.
- OpenAI & Anthropic Claude: Configuring GPT-4o and Claude 3.5 Sonnet cloud endpoints.
- Local Offline Ollama: Running open-source models (like Llama 3 or Mistral) locally on your GPU for a fully offline workflow.
[!WARNING] Keep your API keys private. Never commit your
ProjectSettingsfolder or files containing raw API keys to public repositories. Decnet stores API credentials locally inside Unity’sEditorPrefs(which are unique to your local machine registry), not inside project asset files.