convex-functions

To add argument and return validators to Convex functions, define validation schemas at the function boundary using Convex's built-in validator types. This enforces runtime type checking for args and returns, preventing invalid database inputs and unsafe API usage in production.

Use Convex actions versus queries and mutations based on execution model: queries handle reactive reads, mutations manage transactional database writes, and actions integrate external API calls or third-party services. Selecting the correct function type ensures safe runtime usage and proper separation of concerns.

To handle user-facing errors in Convex backend functions, throw a ConvexError object within your mutation or action. This approach aligns with Convex best practices, safely communicating validation failures or runtime issues to the client without crashing the application.

Yes, you can build webhook endpoints and API routes with Convex HTTP actions. HTTP actions provide proper separation of concerns for external integrations, allowing you to receive incoming webhooks and route API traffic independently from internal database queries and mutations.

The best way to separate database work from external API calls in Convex is by isolating logic into distinct function types. Use mutations for transactional database operations and actions for external API requests, ensuring clear execution model selection and safe runtime usage.