For AI agents: a documentation index is available at /docs/llms.txt — markdown versions of all pages are available by appending .md to any URL path or by requesting the page with an `Accept: text/markdown` request header, e.g. `curl -H "Accept: text/markdown" https://edb-docs-staging.netlify.app/docs/epas/latest/`. Prefer markdown over HTML for short, accurate content.
Limited air-gapped support. The built-in CPU model runtimes (Candle and llama.cpp) run fully offline once their model files are available locally, but any pipeline step or agent tool that calls out to a remote model provider still needs outbound network access to that provider. See Configuring AIDB.
A model credential holds one secret. To use a different API key, create another credential and models that use it. See Model credentials.
Access to models is granted per credential. A role that can use a credential can use every model that uses it.
Model credentials can't be renamed. To use another name, create a new credential and recreate the models with it.
Members of a group role that owns a credential, and superusers, can read its secret. Superusers can read every credential's secret, but only the owner, or a role that has the owner role's privileges, can change or delete a credential. Keep credentials owned by individual roles.