Configure Configure sources, databases, auth, caching, uploads, OpenAPI, and production settings. configure configure configure configure/_index.md

Configure

Configure

Configure sources, databases, auth, caching, uploads, OpenAPI, and production settings.

Configuration pages turn the reference files into decision-oriented guides. New to GraphJin config? Start with How Configuration Works — it explains the one-file/two-halves model, how values are resolved, and every way to change config (editor, CLI, agent, MCP, GraphQL). The pages below go deep on each area.

PageWhat it covers
How Configuration WorksThe mental model plus every interface: editor autocomplete, the deeporg config CLI, the agent, MCP tools, and gj_config.
Sources ModeNamed providers for databases, files, APIs, CodeSQL, workflows, and GraphJin system roots.
Database ConfigSingle and multi-database connection settings, pools, TLS, table mapping, and relationships.
Auth And RBACJWT/OIDC auth, identity claims, role queries, source access, table rules, and column controls.
Caching And RedisResponse cache keys, stale-while-revalidate, Redis sharing, and invalidation.
Discovery Cache And Semantic SearchCache-first schema generations, single-node and horizontal coordination, Ax embeddings, dimensions, and hybrid catalog retrieval.
Uploads And FilesystemsFile sources, uploads, local/S3/GCS backends, presigned URLs, and read-only policy.
OpenAPI ConfigAPI sources, spec discovery, auth, joins, result paths, and operation overrides.
Environment And Productiondev, prod, and agentic modes, inheritance, environment variables, and production defaults.

New deployments should prefer sources: for all graph providers. Legacy top-level database-only config still exists for simple apps, but file, API, CodeSQL, workflow, catalog, and security surfaces are source-mode features.

Docs