GraphQL and REST Together in Ledger1

Ledger1 pairs GraphQL for product-facing experiences with REST for external system integrations. This keeps the product graph clean while respecting the operational realities of vendors.

When to use GraphQL

• Product UI that composes data from multiple domains.
• Strong typing and introspection for fast developer iteration.
• Co-located queries with components to avoid over-fetching.

When to use REST

• Integrations with third-party systems (POS, payroll, inventory) that expose REST.
• Webhooks and idempotent task endpoints.
• Long-running jobs (export/import) with job status polling.

Patterns that work

• GraphQL behind the app; REST at the boundary for vendors.
• Error shaping: convert vendor-specific errors to stable domain errors.
• Versioning: REST endpoints versioned; GraphQL evolved with deprecation.
• Caching: REST responses cached; GraphQL results memoized on client.

In the Varuni demo

• Apollo Client with typed operations for dashboard queries.
• REST routes for Toast ingestion, backfill, and reconciliation jobs.
• Shared auth utilities and consistent request tracing across both.

This hybrid approach keeps the developer experience excellent while isolating vendor volatility from the product graph.