Configuration¶

Configuration comes from src/backend/.env locally or deployment environment variables. Example values show shape only; never commit real secrets or passwords.

Core and security¶

Variable	Purpose
`SECRET_KEY`, `JWT_SECRET_KEY`	Encryption and JWT signing; production must use distinct, high-entropy values.
`JWT_ALGORITHM`, `JWT_EXPIRE_MINUTES`	JWT algorithm and expiry.
`HOST`, `PORT`, `CORS_ORIGINS`	Listener and allowed origins; explicitly use `HOST=0.0.0.0` in containers.
`ADMIN_*`	Initial-admin bootstrap; replace defaults and restrict use in production.

Database and data¶

Variable	Purpose
`DATABASE_TYPE`, `DATABASE_URL`	Application database (SQLite/PostgreSQL/MySQL).
`AKSHARE_DATA_DATABASE_URL`	Separate AkShare/MySQL market-data warehouse.
`AKSHARE_SCHEDULER_*`, `AKSHARE_SCRIPT_ROOT`	Data-script and scheduler configuration.
`SYNC_LOCAL_MYSQL_*`	Local MySQL connection for data synchronization; configure only when syncing.
`DB_AUTO_CREATE_SCHEMA`, `DB_AUTO_CREATE_DEFAULT_ADMIN`	Development bootstrap flags; usually off in production.

AI and RAG¶

Variable	Purpose
`AI_CHAT_ENABLED`, `AI_CHAT_BASE_URL`, `AI_CHAT_API_KEY`, `AI_CHAT_MODEL`	OpenAI-compatible generation provider.
`AI_CHAT_TIMEOUT`, `AI_CHAT_TEMPERATURE`, `AI_CHAT_MAX_TOKENS`	Invocation limits and sampling settings.
`RAG_VECTOR_ENABLED`	Enables local semantic-vector retrieval.
`RAG_EMBEDDING_MODEL`, `RAG_VECTOR_COLLECTION`, `RAG_VECTOR_UPSERT_BATCH_SIZE`	Embedding model, collection, and batching.

Without a configured model, the knowledge base can still provide index and lexical-retrieval diagnostics; do not present an empty API key as an available provider. Install the rag extra before enabling vector retrieval.

Minimal example¶

DATABASE_TYPE=sqlite
DATABASE_URL=sqlite+aiosqlite:///../../data/dev/backtrader.db
SECRET_KEY=replace-me
JWT_SECRET_KEY=replace-me-with-another-random-value
AI_CHAT_ENABLED=false
RAG_VECTOR_ENABLED=true

See Production for deployment configuration, backup, and access controls.