asn1-per-ts

Try it in the sandbox

TypeScript library for encoding and decoding data using ASN.1 PER (Packed Encoding Rules) unaligned variant (ITU-T X.691).

Features

• Bit-level buffer with MSB-first encoding and automatic growth
• Primitive codecs: BOOLEAN, INTEGER, ENUMERATED, BIT STRING, OCTET STRING, OBJECT IDENTIFIER, IA5String, VisibleString, UTF8String, NULL
• Composite codecs: CHOICE, SEQUENCE, SEQUENCE OF
• Constraint support: value ranges, size constraints, extensibility markers, default values
• Schema-driven encoding: define types as JSON, encode/decode plain objects
• Typed from the schema: decode() returns the type the schema describes and encode() rejects values that do not fit — at compile time, with no runtime validation and no extra dependency
• Three ways to say it: an inline SchemaNode, the asn builder DSL, or TypeScript generated from a .asn file — all sharing one interchange format and one codec interface
• Metadata decoding: decodeWithMetadata returns a tree of DecodedNode objects with bit positions, raw bytes, and codec references for every field
• Pre-encoded passthrough: embed already-encoded bits verbatim with RawBytes, via codec.raw

Install

npm install asn1-per-ts

The package also installs an asn1-per-ts command:

npx asn1-per-ts types  ticket.asn src/generated/ticket.ts  # TypeScript types + codecs
npx asn1-per-ts schema ticket.asn ticket.schema.json       # SchemaNode JSON

Usage

Low-level codec API

import { BitBuffer, IntegerCodec, BooleanCodec, SequenceCodec } from 'asn1-per-ts';

// Constrained integer (0..255) uses 8 bits
const intCodec = new IntegerCodec({ min: 0, max: 255 });

const buf = BitBuffer.alloc();
intCodec.encode(buf, 42);
buf.reset();
console.log(intCodec.decode(buf)); // 42

Schema-driven API

import { SchemaCodec } from 'asn1-per-ts';

const codec = new SchemaCodec({
  type: 'SEQUENCE',
  fields: [
    { name: 'id', schema: { type: 'INTEGER', min: 0, max: 255 } },
    { name: 'active', schema: { type: 'BOOLEAN' } },
    { name: 'status', schema: { type: 'ENUMERATED', values: ['pending', 'approved', 'rejected'] } },
  ],
});

const hex = codec.encodeToHex({ id: 42, active: true, status: 'approved' });
console.log(hex);

const decoded = codec.decodeFromHex(hex);
console.log(decoded);

Types come from the schema

The schema is read as a literal type, so the decoded type is derived from it — no interface to write, no cast to make:

const decoded = codec.decodeFromHex(hex);
//    ^? { id: number; active: boolean; status: 'pending' | 'approved' | 'rejected' }

decoded.status; // 'pending' | 'approved' | 'rejected'

codec.encodeToHex({ id: 42, active: true, status: 'unknown' });
//                                                ~~~~~~~~~ not one of the ENUMERATED values

OPTIONAL fields become optional properties, DEFAULT fields are always present after decoding, CHOICE becomes a discriminated union, and $ref resolves through createCodecs. When the schema is only known at runtime (parsed from JSON or from ASN.1 text), everything degrades to unknown exactly as before.

Three front-ends, one interchange format

SchemaNode is the interchange format; the three ways of describing a type all produce it and all hand back the same TypedCodec.

import { SchemaCodec, asn } from 'asn1-per-ts';

// 1. inline SchemaNode — the schema is data, types are read from the literal
const fromLiteral = new SchemaCodec({ type: 'SEQUENCE', fields: [/* … */] });

// 2. the asn DSL — types ride in the value, no widening traps, short errors
const fromDsl = asn.codec(
  asn.sequence({
    id: asn.integer({ min: 0, max: 255 }),
    status: asn.enumerated(['pending', 'approved']),
    nickname: asn.ia5String().optional(),
  }),
);

// 3. generated from ASN.1 — named types, cheapest to compile, best errors
//    npx asn1-per-ts types ticket.asn src/generated/ticket.ts
import { codecs } from './generated/ticket.js';

Pick by where the schema comes from: generate from .asn files, use the DSL for hand-written schemas, use SchemaNode when the schema itself is data.

• examples/typed-api.md — inline schemas and the Infer types
• examples/dsl.md — the asn builder DSL
• examples/codegen.md — generating TypeScript from ASN.1

Decoding with Metadata

decodeWithMetadata returns a DecodedNode tree with full encoding metadata (bit offsets, bit lengths, raw bytes, codec references) for every field. Use stripMetadata to convert back to a plain object identical to decode().

import { SchemaCodec, stripMetadata } from 'asn1-per-ts';

const codec = new SchemaCodec({
  type: 'SEQUENCE',
  fields: [
    { name: 'id', schema: { type: 'INTEGER', min: 0, max: 255 } },
    { name: 'active', schema: { type: 'BOOLEAN' } },
  ],
});

const hex = codec.encodeToHex({ id: 42, active: true });
const node = codec.decodeFromHexWithMetadata(hex);

// The tree is typed from the schema — no cast needed
console.log(node.value.id.value);           // 42 (number)
console.log(node.value.id.meta.bitOffset);  // 0
console.log(node.value.id.meta.bitLength);  // 8
console.log(node.value.id.meta.rawBytes);   // Uint8Array([0x2a])

// Strip metadata to get the plain object, with the same type as decode()
const plain = stripMetadata(node);
// plain === { id: 42, active: true }

Extension Markers

Extension markers (...) indicate that a type may be extended in future versions, providing forward compatibility. When present, PER encoding includes a 1-bit extension marker prefix (0 = not extended, 1 = extended).

In ASN.1 notation

Use ... to mark a type as extensible in ASN.1 schema text parsed by the built-in parser:

-- SEQUENCE: extension marker separates root fields from extension additions
MessageV1 ::= SEQUENCE {
    id     INTEGER (0..255),
    name   IA5String,
    ...                        -- extensible, no extensions yet
}

MessageV2 ::= SEQUENCE {
    id     INTEGER (0..255),
    name   IA5String,
    ...,                       -- extension marker
    email  IA5String           -- extension addition
}

-- ENUMERATED: extension marker separates root values from extension values
Color ::= ENUMERATED { red, green, blue, ... }
ColorV2 ::= ENUMERATED { red, green, blue, ..., yellow, purple }

-- CHOICE: extension marker separates root alternatives from extension alternatives
Shape ::= CHOICE { circle BOOLEAN, ..., polygon INTEGER }

-- Constraints: extensible constraints allow values outside the root range
FlexInt ::= INTEGER (0..100, ...)
FlexStr ::= OCTET STRING (SIZE (1..50, ...))

In JSON SchemaNode definitions

When building schemas directly as JSON SchemaNode objects, indicate extensibility with these properties:

// SEQUENCE: provide extensionFields (even empty [] to mark as extensible)
const schema: SchemaNode = {
  type: 'SEQUENCE',
  fields: [
    { name: 'id', schema: { type: 'INTEGER', min: 0, max: 255 } },
  ],
  extensionFields: [],  // extensible with no additions yet
};

// ENUMERATED: provide extensionValues
{ type: 'ENUMERATED', values: ['red', 'green'], extensionValues: [] }

// CHOICE: provide extensionAlternatives
{ type: 'CHOICE', alternatives: [...], extensionAlternatives: [] }

// Constrained types: set extensible: true
{ type: 'INTEGER', min: 0, max: 100, extensible: true }
{ type: 'BIT STRING', minSize: 1, maxSize: 50, extensible: true }

Key distinction: providing an empty array (extensionFields: []) marks the type as extensible, while omitting the property entirely (extensionFields: undefined) means the type is not extensible. This matters because extensible types include the 1-bit extension marker in the encoding.

Supported Types

| Type | Description | |------|-------------| | BOOLEAN | Single bit | | INTEGER | Constrained, semi-constrained, or unconstrained | | ENUMERATED | Indexed enumeration with optional extensions | | BIT STRING | Bit sequences with size constraints | | OCTET STRING | Byte sequences with size constraints | | OBJECT IDENTIFIER | Dot-notation OIDs | | IA5String | ASCII strings with optional alphabet constraints | | VisibleString | Printable strings with optional alphabet constraints | | UTF8String | UTF-8 encoded strings | | NULL | Zero-bit placeholder | | CHOICE | Tagged union of alternatives | | SEQUENCE | Ordered fields with OPTIONAL/DEFAULT support | | SEQUENCE OF | Homogeneous list with size constraints |

Examples

The examples/ directory contains detailed usage guides with code samples:

• Typed API - Types inferred from an inline schema: Infer, InferInput, InferMetadata, $ref registries
• Builder DSL - The asn DSL, where types ride in the value instead of being read from literals
• Code generation - Generating TypeScript types and codecs from a .asn file
• Schema Parser - Parse ASN.1 text notation into SchemaNode definitions, constraint options, extension markers, CLI usage
• Encoding - Encode JavaScript objects to PER unaligned binary using SchemaCodec or low-level codecs
• Decoding - Decode PER unaligned binary data back into objects

Sister Project

dosipas-ts — a TypeScript library built on top of asn1-per-ts for encoding and decoding DOSIPAS / ERA electronic ticket data.

Development

npm test             # Run tests
npm run build        # Build library to dist/
npx tsc --noEmit     # Type-check without emitting

# Run the CLI from source, without building
npx tsx src/cli/main.ts types <in.asn> [out.ts]

Changes are recorded in CHANGELOG.md; releasing is described in PUBLISH.md.

Website

The website/ directory contains a React + TailwindCSS demo app for interactive ASN.1 PER encoding and decoding.

cd website
npm install
npm run build     # Build for GitHub Pages (uses ./ asset path)
npm run dev       # Development server

License

MIT