FAQ

Is this production-ready?
HumeSQL provides guardrails and uses parameterized execution, but you should review outputs, apply least-privilege DB access, and add your own auditing/limits for production environments.

Which databases are supported?
MySQL and MariaDB via mysql-connector-python. Other engines would require additional adapters.

Can I disable schema caching?
Yes: cache_schema=False in Python or --no-cache-schema via CLI. You can also call refresh_schema_cache() after schema changes.

How do I select a different Gemini model?
Pass model="gemini-1.5-pro-latest" (Python) or --model gemini-1.5-pro-latest (CLI). Ensure your API key has access.

How are API keys resolved?
Order: api_key argument > HUMESQL_AI_KEY > HUMANSQL_AI_KEY (legacy) > GEMINI_API_KEY.