What does the parser understand?

It reads model blocks, field names, scalar and model types, list (Type[]) and optional (Type?) markers, and inline attributes such as @id, @unique, @default(...), @relation(...) and @updatedAt. enum and datasource blocks are recognised and skipped from the model view.

How are relations detected?

A field whose type matches another model name is treated as a relation. The tool marks it one-to-many when the type is a list (Model[]) and one-to-one or many-to-one for a single Model reference, and surfaces any @relation foreign-key fields.

Does it validate the schema?

It is a viewer, not a validator. It will parse well-formed schemas and ignore syntax it doesn't recognise rather than reporting Prisma validation errors. Use the Prisma CLI for strict validation.

Can I get JSON out of it?

Yes. Alongside the readable tables, the tool produces a structured JSON object describing each model and its fields, which you can copy for documentation, code generation or diagrams.

Is my schema sent anywhere?

No. Tokenising and rendering happen entirely in your browser, so model names, field names and any embedded comments never leave your device.

Prisma Schema Viewer

A Prisma schema viewer turns the declarative schema.prisma file into a readable map of your data model — every model, its fields and types, the attributes that constrain them, and how models relate to one another. It is handy for onboarding, documentation and quickly understanding an unfamiliar codebase, and it runs locally so private schemas stay private.

How it works

The tool tokenises the schema rather than using a regex on whole lines, so it handles nested attribute parentheses correctly. The steps are:

Strip // and /// comments and locate each top-level block: model, enum, datasource, generator.
For each model Name { … } block, split the body into field lines.
Parse each field as name Type[modifier] @attr @attr(...) where the modifier is ? (optional) or [] (list).
Classify the type: a built-in scalar (String, Int, Boolean, DateTime, Float, Decimal, BigInt, Json, Bytes), an enum, or another model name (a relation).

A field is a relation when its type equals a model name. A list type (Post[]) implies a one-to-many side; a singular model reference implies one-to-one or many-to-one. The @relation(fields: […], references: […]) attribute pinpoints the foreign-key column.

Tips and notes

Optional fields (Type?) are marked nullable; list fields (Type[]) are arrays.
@id, @unique, @default(...), @updatedAt and @map(...) are surfaced per field.
Enum blocks are listed separately so you can see allowed values.
The structured JSON output is convenient for feeding into diagram or documentation generators.