Configuration
Omelette is configured via environment variables. Copy .env.example to .env and customize.
Application
| Variable | Description | Default |
|---|---|---|
APP_ENV | development, production, or testing | development |
APP_DEBUG | Enable debug mode | true |
APP_HOST | Backend bind address | 0.0.0.0 |
APP_PORT | Backend port | 8000 |
Database
| Variable | Description | Default |
|---|---|---|
DATABASE_URL | SQLite connection string | sqlite:///./data/omelette.db |
Data Storage
| Variable | Description |
|---|---|
DATA_DIR | Base path for PDFs, OCR output, ChromaDB |
PDF_DIR | PDF storage (default: {DATA_DIR}/pdfs) |
OCR_OUTPUT_DIR | OCR output (default: {DATA_DIR}/ocr_output) |
CHROMA_DB_DIR | ChromaDB path (default: {DATA_DIR}/chroma_db) |
LLM Provider
Set LLM_PROVIDER to one of: openai, anthropic, aliyun, volcengine, ollama, mock.
OpenAI
| Variable | Description |
|---|---|
OPENAI_API_KEY | OpenAI API key |
OPENAI_MODEL | Model name (default: gpt-4o-mini) |
Anthropic
| Variable | Description |
|---|---|
ANTHROPIC_API_KEY | Anthropic API key |
ANTHROPIC_MODEL | Model name (default: claude-sonnet-4-20250514) |
Aliyun Bailian
| Variable | Description |
|---|---|
ALIYUN_API_KEY | Aliyun API key |
ALIYUN_BASE_URL | OpenAI-compatible endpoint |
ALIYUN_MODEL | Model name (e.g. qwen3.5-plus) |
Volcengine Doubao
| Variable | Description |
|---|---|
VOLCENGINE_API_KEY | Volcengine API key |
VOLCENGINE_BASE_URL | OpenAI-compatible endpoint |
VOLCENGINE_MODEL | Model name |
Ollama (local)
| Variable | Description |
|---|---|
OLLAMA_BASE_URL | Ollama server URL (default: http://localhost:11434) |
OLLAMA_MODEL | Model name (default: llama3) |
Mock
Use LLM_PROVIDER=mock for testing without API keys. No additional variables required.
Embeddings
| Variable | Description | Default |
|---|---|---|
EMBEDDING_PROVIDER | local, api, or mock | local |
EMBEDDING_MODEL | Model name (local: HuggingFace; api: OpenAI-compatible) | BAAI/bge-m3 |
- local: Uses sentence-transformers with GPU auto-detection
- api: Uses OpenAI-compatible embedding API
- mock: Deterministic mock for tests
GPU
| Variable | Description | Default |
|---|---|---|
CUDA_VISIBLE_DEVICES | Comma-separated GPU IDs for OCR/embeddings | 0,3 |
Proxy
| Variable | Description |
|---|---|
HTTP_PROXY | HTTP proxy URL |
HTTPS_PROXY | HTTPS proxy URL |
Example: HTTP_PROXY=http://127.0.0.1:20171/
External APIs
| Variable | Description |
|---|---|
SEMANTIC_SCHOLAR_API_KEY | Optional; increases Semantic Scholar rate limit |
UNPAYWALL_EMAIL | Required for Unpaywall PDF lookup |
Frontend Settings
LLM provider, model, temperature, and API keys can be configured via the Settings page at /settings in the web UI. These settings override environment variables for the current user and are stored in the database. Use this for per-user customization without editing .env.