AI Setup
LearnHouse includes AI-powered features for learning assistance and content editing. The primary AI provider is Google Gemini, with OpenAI also supported as an alternative. These features are optional and disabled by default.
LearnHouse uses llama-index with pgvector for RAG (Retrieval-Augmented Generation), enabling AI assistants to answer questions using your course content as context.
Enabling AI
Using Google Gemini (Recommended)
Set the following environment variables in your .env file:
LEARNHOUSE_GEMINI_API_KEY=your-gemini-api-key
LEARNHOUSE_IS_AI_ENABLED=trueTo get a Gemini API key, visit Google AI Studio .
Using OpenAI
Alternatively, you can use OpenAI as your AI provider:
LEARNHOUSE_OPENAI_API_KEY=sk-your-openai-api-key
LEARNHOUSE_IS_AI_ENABLED=trueAfter updating your .env file, restart your instance:
learnhouse stop
learnhouse startAI features require a valid API key with sufficient credits. Usage costs are billed directly by your chosen provider based on the models and volume used.
Per-Organization AI Configuration
AI settings can also be configured on a per-organization basis through the organization configuration in the admin panel. This allows you to:
- Enable or disable AI for specific organizations
- Select which AI model to use per organization
- Control which AI features are available
AI Feature Toggles
The following AI features can be individually enabled or disabled per organization:
| Feature | Description |
|---|---|
activity_ask | AI assistant within learning activities, allows learners to ask questions about activity content |
course_ask | AI assistant at the course level, answers questions using the full course material as context |
editor | AI-powered writing assistance in the content editor |
global_ai_ask | Global AI assistant available across the platform |
Usage Limits
You can configure usage limits per organization to control API costs. Limits can be set on the number of AI requests per user per day, preventing unexpected charges on your provider account.
Model Selection
The AI model used for generating responses can be configured per organization. This allows you to balance quality and cost by choosing different models for different use cases.
