Convex Schema Builder

Convex is a reactive backend platform for building full-stack applications. It provides a real-time database, serverless functions, and automatic type safety — all in one package. The schema defined in convex/schema.ts enforces your data model and generates TypeScript types automatically.

defineSchema wraps your table definitions and tells Convex the shape of your data. Each table is defined with defineTable() and a validator object using the v.* validators. Convex uses this to enforce data integrity and generate TypeScript types for your queries and mutations.

Convex provides: v.string() for text, v.number() for numbers, v.boolean() for booleans, v.id('tableName') for references to other tables, v.array(v.string()) for arrays, v.object({...}) for nested objects, v.optional(v.string()) for optional fields, v.null() for null values, and v.union(...) for union types.

v.id('tableName') creates a reference to another table's document. It stores the document's unique ID (_id field) which Convex uses to join data across tables. For example, v.id('users') on a posts table means each post references a user document. Use db.get(post.authorId) in a query to fetch the referenced document.

Add .index('by_fieldName', ['fieldName']) to a defineTable() call to create a searchable index. Convex indexes are used in db.query('tableName').withIndex('by_fieldName', q => q.eq('fieldName', value)). Compound indexes on multiple fields support range queries. Indexes are required for efficient filtered queries — full table scans are not supported in production.

Yes. convex/schema.ts should always be committed to your repository. It is the source of truth for your data model and is required for Convex to enforce types and generate TypeScript definitions. Running npx convex dev will validate your schema against the deployed state and warn you of breaking changes.