remote-zip

API documentation

Fetch file listings and individual files from a remote ZIP file.

Features

Without downloading the entire ZIP:

• Fetch individual files in a remote ZIP (buffered or streaming)
• Fetch file listings
• ZIP64 archives (>4 GiB / >65,535 entries)
• Encrypted entries: traditional ZipCrypto and WinZip AES (AE-1/AE-2)
• CP437 and UTF-8 filenames; optional CRC-32 verification

The gist of what the library does is:

1. Get the content size of the ZIP file using a HTTP HEAD request.
2. Get the last 100 bytes using HTTP Range queries (hoping it's enough) to read the end-of-central-directory (EOCD) section.
3. Using the EOCD, read the central directory (CD) section with a Range request. This section contains a complete file listing and file byte offsets in the ZIP.
4. To get individual files, use another Range request with an offset to get the local file header + compressed data.

Limitations

• PKWare strong encryption (general-purpose bit 6) is detected and rejected
• Multi-disk / split archives are not supported

Decrypting WinZip AES entries uses the Web Crypto API (crypto.subtle), which in browsers is only available in a secure context (HTTPS or localhost). Traditional ZipCrypto works everywhere but is cryptographically weak.

Install

npm install @gyng/remote-zip
Copy

yarn add @gyng/remote-zip
Copy

Usage

See the generated API documentation.

Server requirements (CORS & Range)

The remote server must support HTTP Range requests (respond 206 Partial Content). When used cross-origin from a browser, it must also send the right CORS headers. The browser issues the CORS preflight (OPTIONS) automatically — there is nothing to configure on the client — but the server needs to allow it:

• Access-Control-Allow-Origin: <your origin>
• Access-Control-Allow-Methods: GET, HEAD (add POST if you set a custom method)
• Access-Control-Allow-Headers: Range (plus Authorization / any custom headers you pass)
• Access-Control-Expose-Headers: Content-Length, Content-Range — required, or the browser hides those response headers and populate() cannot read the archive size

Range is not a CORS-safelisted request header, so any cross-origin request triggers a preflight. Static hosts like S3, GitHub Pages, and nginx support Range out of the box; you only need to configure the CORS headers above.

Basic

const url = new URL("http://www.example.com/test.zip");
const remoteZip = await new RemoteZipPointer({ url }).populate();
const fileListing = remoteZip.files(); // RemoteZipFile[]
const uncompressedBytes = await remoteZip.fetch("test.txt"); // Uint8Array
Copy

Streaming

For large entries, fetchStream returns a ReadableStream<Uint8Array> of the uncompressed bytes so you can process them incrementally without buffering the whole file. maxUncompressedSize, if set, is enforced mid-stream.

const stream = await remoteZip.fetchStream("big.bin");
for await (const chunk of stream) {
  // handle each chunk
}
Copy

With more features

const additionalHeaders = new Headers();
additionalHeaders.append("X-Example", "foobar");
const url = new URL("http://www.example.com/test.zip");

const remoteZip = await new RemoteZipPointer({
  url,
  additionalHeaders,
  method: "POST",
  credentials: "include",
  // New request options (all optional), applied to every request:
  redirect: "error", // avoid leaking auth headers cross-origin on a 30x
  timeoutMs: 10_000, // per-request timeout
  signal: AbortSignal.timeout(30_000), // or your own AbortController signal
  requestInit: { cache: "no-store" }, // escape hatch merged into every fetch
}).populate();

// Guard untrusted archives against decompression bombs, and pass a per-call
// signal/timeout if you like:
const uncompressedBytes = await remoteZip.fetch("test.txt", additionalHeaders, {
  maxUncompressedSize: 50 * 1024 * 1024,
  verifyCrc: true, // check the decompressed bytes against the entry's CRC-32
  password: "hunter2", // for ZipCrypto / WinZip AES encrypted entries
});
Copy

Dev

npm ci            # install (reproducible, honours .npmrc cooldown)
npm run d         # watch and build
npm run t:watch   # watch and test
npm run lint      # prettier + eslint
npm run typecheck # tsc --noEmit (also checks tests)
npm run build     # type declarations + esm/cjs bundles
npm test          # run the test suite once
Copy

License

Licensed under either of

• Apache License, Version 2.0 (LICENSE-APACHE)
• MIT license (LICENSE-MIT)

at your option.

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.