ASN1js

Abstract Syntax Notation One (ASN.1) is a standard and notation that describes rules and structures for representing, encoding, transmitting, and decoding data in telecommunications and computer networking. ASN1js is a pure JavaScript library implementing this standard. ASN.1 is the basis of all X.509 related data structures and numerous other protocols used on the web.

Install

npm install asn1js

Documentation, the API reference, and runnable examples are at asn1js.org.

Breaking Changes In v4

v4 is ESM-only and requires Node.js ^20.19.0 || >=22.12.0 (was >=12.0.0). CommonJS code can still require("asn1js") on these versions; Node.js 22.0–22.11 can't. The exported API is unchanged, but some encoding and validation behavior changed (last three items).

• TypeScript compiling to CommonJS: use "module": "node20" or "nodenext" (node16 fails with TS1479). ESM and bundler setups need no changes.
• Jest in CommonJS mode: switch to ESM mode or Vitest, or configure an ESM-to-CommonJS transformer and make sure transformIgnorePatterns doesn't exclude asn1js and its ESM dependencies, such as @peculiar/utils.
• Deep imports (asn1js/build/index.js, asn1js/build/index.es.js) are blocked by exports; import from asn1js.
• One asn1js per app: libraries such as pkijs and @peculiar/asn1-schema check instanceof asn1js.Sequence and similar, so objects from asn1js 4 fail those checks in a library that still depends on asn1js 3. Upgrade them together.
• New RangeErrors: new Integer({ value }) and the valueDec setter reject non-integer numbers (v3 truncated 1.5 to 1 and encoded NaN as empty). Encoding a UTCTime with a year outside 1950–2049 throws, including new UTCTime({ valueDate }); v3 wrapped 2050 to 1950. Use GeneralizedTime for those dates.
• Encodings that differ from v3: Integer.fromBigInt encodes negative values correctly (v3: -32767n → 0x81, -128n → 0xff80). UniversalString writes characters outside the BMP as one UCS-4 code point (U+1F600 → 0001f600, v3 wrote a surrogate pair; v3 output still decodes). OID arcs and tag numbers from 2^49 to 2^53 now encode (v3 produced an empty value).
• Decoding: a BMPString with an odd length or a UniversalString whose length isn't a multiple of 4 returns an error (offset: -1) instead of throwing. UniversalString lone surrogates and code points above U+10FFFF become U+FFFD with a warning. OID arcs and tag numbers above 2^53 stay hex-only and re-encode unchanged. "Needlessly long encoded length" now checks the first length octet, so lengths such as 83 00 00 07 get the warning.

Introduction

ASN1js is the first library for BER encoding/decoding in JavaScript designed for browser use. BER is the basic encoding rules for ASN.1 that all others are based on, DER is the encoding rules used by PKI applications - it is a subset of BER. The ASN1js library was tested against freely available ASN.1:2008 test suite, with some limitations related to JavaScript language.

Features of the library

• Based on latest features of JavaScript language from ES2015 standard;
• ASN1js is a "base layer" for full-featured JS library PKIjs, which is using Web Cryptography API and has all classes, necessary to work with PKI-related data;
• Fully object-oriented library. Inheritance is used everywhere inside the lib;
• Working with HTML5 data objects (ArrayBuffer, Uint8Array etc.);
• Working with all ASN.1:2008 types;
• Working with BER encoded data;
• All types inside the library constantly stores information about all ASN.1 sub blocks (tag block, length block or value block);
• User may have access to any byte inside any ASN.1 sub-block;
• Any sub-block may have unlimited length, as it described in ASN.1 standard (even "tag block");
• Ability to work with ASN.1 string data types (including all "international" strings like UniversalString, BMPString, UTF8String) by passing native JavaScript strings into constructors. And vice versa - all initially parsed data of ASN.1 string types right after decoding automatically converts into native JavaScript strings;
• Same with ASN.1 date-time types: for major types like UTCTime and GeneralizedTime there are automatic conversion between "JS date type - ASN.1 date-time type" + vice versa;
• Same with ASN.1 OBJECT-IDENTIFIER (OID) data-type: you can initialize OID by JavaScript string and can get its string representation by calling oid.getValue();
• Working with "easy-to-understand" ASN.1 schemas (pre-defined or built by user);
• Has special types to work with ASN.1 schemas:
  • Any
  • Choice
  • Repeated
• User can name any block inside ASN.1 schema and easily get information by name;
• Ability to parse internal data inside primitively encoded data types and automatically validate it against special schema;
• All types inside library are dynamic;
• All types can be initialized in static or dynamic ways.
• ASN1js fully tested against ASN.1:2008 TestSuite.

Examples

Runnable examples are on the website:

• Basics: decode BER and DER, encode to DER, convert to JavaScript values, explicit and implicit tags
• Schemas: validate against a schema, CHOICE and SEQUENCE OF, structures inside an OCTET STRING
• Real-world data: read an X.509 certificate, read a public key, parse untrusted input

For complete PKI structures such as certificates, CMS, and PKCS#10 requests, see PKIjs, which is built on ASN1js.

Object Identifiers

ObjectIdentifier and RelativeObjectIdentifier preserve large decimal components exactly, including UUID-based OIDs such as 2.25.329800735698586629295641978511506172918. getValue(), ASCII output, and JSON values use decimal notation even when a component is stored internally as hex-only.

String values must contain complete non-negative decimal components. Absolute OIDs need at least two components: the first is 0, 1, or 2, and the second is at most 39 when the first is 0 or 1. Leading zeros, surrounding whitespace in each component, a leading +, and negative zero are normalized. Partial numbers, hexadecimal prefixes, and exponent notation are rejected.

By default, invalid strings clear the value and set both error and valueBlock.error without throwing. They no longer encode as another OID or as a valid prefix. For compatibility, toBER() still returns an empty OID (0600 or 0d00) on failure; check the OID's diagnostics before using the output, including when it is nested in a constructed value. Parent blocks do not propagate the child's error. A successful setValue() clears previous value errors. Constructors with an omitted value, undefined, null, or "" create empty placeholders without errors for schema declarations. An explicit setValue("") still records a validation error.

Pass strict: true to make invalid construction, value assignment, and encoding throw an Error. The same decimal input rules and normalizations apply in both modes. An explicit value: "" is invalid in a strict constructor; omit the value to create a strict schema placeholder. RelativeObjectIdentifier supports the same option.

const oid = new asn1js.ObjectIdentifier({
  value: "1.2.840.113549",
  strict: true,
});

oid.setValue("1.40"); // Throws; the object is left invalid with diagnostics.

Strict encoding checks the current SID values on every toBER() call, including toBER(true), so later mutations cannot silently produce an empty value. After a failed assignment or encoding, use a successful setValue() or decode a complete replacement value to recover; changing the SID array or turning strict mode off does not clear recorded errors. Constructors without a value remain valid strict schema placeholders, but encoding an empty placeholder throws. toString("hex") also encodes the value and throws for invalid strict OIDs. isEqual() can throw when it needs to encode an invalid strict OID for comparison.

OID instance and value-block fromBER() calls replace previous SID values, lengths, and value diagnostics. Failed decoding discards any parsed prefix and reports the error without throwing. The instance method expects content octets and a prepared length block; use the top-level fromBER() function to decode a complete BER encoding.

You can also set oid.strict = true on an existing or decoded OID. Changing the flag does not validate or modify its value immediately. The flag defaults to false, is not serialized in BER or JSON, and does not change the permissive behavior of fromBER().

An error from a strict OID aborts the normal encoding of its enclosing Sequence, Set, or Constructed value. Enable strict mode on each OID that needs this behavior. If encoding throws with an explicit ViewWriter, discard that writer: earlier writes are retained and are not rolled back. Strict mode checks OID values; it does not validate arbitrary changes to tag/length blocks or custom serializer implementations.

Parsing Untrusted Input

fromBER() accepts optional parser resource limits for untrusted ASN.1 input.

const asn1 = asn1js.fromBER(inputBuffer, {
  maxDepth: 100,
  maxNodes: 10000,
  maxContentLength: 16 * 1024 * 1024,
});

When any limit is exceeded the parser returns a normal FromBerResult error instead of exhausting the JavaScript stack or continuing to parse excessive input. See the untrusted input example.

By default, primitive OCTET STRING values and primitive BIT STRING values with zero unused bits are also checked for a complete embedded ASN.1 value. Set parseEmbedded: false to keep those values as raw bytes without that heuristic interpretation; constructed children continue to be parsed normally.

Input is copied by default. Set copyInput: false to retain a normalized view over the exact input range without copying. The caller must keep the backing storage unchanged and must not detach it while the parsed tree is in use: mutations are visible through parsed byte views and can make them inconsistent with decoded fields. The entire backing buffer remains retained. This option provides no protection against shared-memory races.

For performance-sensitive callers that can meet these storage requirements, the following example explicitly opts out of copying the input. It also independently disables embedded-value detection with parseEmbedded: false; that option does not change input ownership.

const asn1 = asn1js.fromBER(inputBuffer, {
  maxDepth: 100,
  maxNodes: 10000,
  maxContentLength: 16 * 1024 * 1024,
  parseEmbedded: false,
  copyInput: false,
});

Related source code

• C++ ASN1:2008 BER coder/decoder - the "father" of ASN1js project;
• Freely available ASN.1:2008 test suite - the suite which can help you to validate (and better understand) any ASN.1 coder/decoder;
• NPM package for ASN.1:2008 test suite

Suitability

There are several commercial products, enterprise solutions as well as open source project based on versions of ASN1js. You should, however, do your own code and security review before using any open source library in a production application to ensure it will meet your needs.

Contributing

See CONTRIBUTING.md for development setup, commit conventions and the release flow.

License

asn1js is released under the BSD 3-Clause License.