Cursor Plugin Convex Rule Schema Design

Schema Design Best Practices

Design schemas to be document-relational: relatively flat documents with relationships via IDs, not deeply nested structures.

Key Principles

1. Keep documents flat: Avoid deeply nested arrays of objects
2. Use relationships: Link documents via IDs across tables
3. Add indexes early: Index foreign keys (userId, teamId) from the start
4. Limit array sizes: Arrays are capped at 8,192 items—only use when there's a natural limit

Bad: Deep Nesting

// ❌ Don't do this
export default defineSchema({
  users: defineTable({
    name: v.string(),
    posts: v.array(v.object({
      title: v.string(),
      content: v.string(),
      comments: v.array(v.object({
        text: v.string(),
        author: v.string(),
      })),
    })),
  }),
});

This makes it hard to update specific posts, limits you to 8,192 posts per user, and prevents efficient queries.

Good: Relational Design

// ✅ Do this
export default defineSchema({
  users: defineTable({
    name: v.string(),
    email: v.string(),
  }).index("by_email", ["email"]),

  posts: defineTable({
    userId: v.id("users"),
    title: v.string(),
    content: v.string(),
  }).index("by_user", ["userId"])
    .index("by_user_and_created", ["userId", "createdAt"]),

  comments: defineTable({
    postId: v.id("posts"),
    userId: v.id("users"),
    text: v.string(),
  }).index("by_post", ["postId"])
    .index("by_user", ["userId"]),
});

When Arrays Are OK

Arrays work well for:

• Small, bounded collections (e.g., roles, tags)
• Data that's always loaded together
• Natural limits (e.g., max 5 favorites)

users: defineTable({
  name: v.string(),
  roles: v.array(v.union(v.literal("admin"), v.literal("editor"), v.literal("viewer"))),
  favoriteColors: v.array(v.string()), // Small list
}),

Index Your Relationships

Always add indexes for foreign key lookups:

.index("by_user", ["userId"])
.index("by_team", ["teamId"])
.index("by_parent", ["parentId"])