tiny predator at
the app's edge

A JavaScript runtime schema validator written in TypeScript, with type safety and inference.

Documentation
Benchmarks
Architecture
GitHub
dot

Full Boundary Validation

Fluent schemas with constraints, refinements, preparation, casting, composition, and TypeScript inference.

Immutable Schemas

Every fluent change returns a new schema, leaving the original safe to reuse, extend, and share.

Copy-on-Write Semantics

Dot separates validation from mutation. Containers only allocate new results when the validated value differs from the input.

Cheap Schema Construction

Schema construction is inexpensive enough to organize validation around meaning instead of validator overhead.

Optimized Hot Paths

Validation runs through profiled, direct paths designed to minimize allocation and runtime overhead.

Tiny Bundles

Zero dependencies, no code generation, and realistic bundles around 4–6 KB gzipped depending on API usage.

Hi, my name is Dot.

I am small and fast. I do one thing well: I validate stuff.

At the app's edge, a shape comes in, Dot checks it, and you move on to business logic.

ts
import {dot} from '@sajajs/dot'

const UserSchema = dot.object({
    id: dot.string()
        .min(4)
        .as(s => dot.int().gte(1000).validate(Number(s))), // cheap, no hoisting required
    name: dot.string()
        .min(1)
        .prepare(s => s.trim())
        .refine(s => s.toLowerCase() !== 'root', 'reserved name'),
    role: dot.enum(['admin', 'predator']).optional(),
    tags: dot.array(dot.string().min(2)).min(1).nullable()
}).strict()

type User = dot.infer<typeof UserSchema>
// {id: number, name: string, role?: 'admin' | 'predator', tags: string[] | null}

const user: User = UserSchema.validate({
    id: '12354',
    name: ' Dot ',
    role: 'predator',
    tags: ['small', 'fast']
})
// {id: 12354, name: 'Dot', role: 'predator', tags: ['small', 'fast']}