Manage Azure Blob Storage Data

@azure/storage-blob is a TypeScript/JavaScript SDK for Azure Blob Storage - upload, download, list, and manage blobs and containers - working in both Node.js (18+) and browsers, with a client hierarchy from BlobServiceClient (account) to ContainerClient to BlobClient/BlockBlobClient/AppendBlobClient/PageBlobClient. Authentication supports DefaultAzureCredential (recommended), a connection string, StorageSharedKeyCredential (Node.js only), or a SAS token appended to the account URL.

Containers can be created (optionally idempotently with createIfNotExists), listed (optionally prefix-filtered), and deleted. Blob uploads vary by environment and size: a simple string/Buffer upload works everywhere; uploadFile() and uploadStream() (with a buffer size, concurrency, and progress callback) are Node.js-only; browser code uploads a File/Blob/ArrayBuffer via uploadData(). Downloads mirror this - a readable stream everywhere, or Node-only downloadToFile()/downloadToBuffer() convenience methods. Blobs can be listed flat or by hierarchy (treating /-delimited names as virtual directories), deleted (optionally including snapshots), and server-side copied via beginCopyFromURL() as a pollable long-running operation. Properties (content type, length, last-modified, ETag), custom metadata, and HTTP headers (content type, cache control, content disposition) can all be read or set independently.

SAS tokens can be generated for a blob, a container, or the whole account, each with its own permission set (e.g. read-only for a blob SAS, or racwdl for a container SAS) and an explicit expiry - SAS generation is Node.js-only since it requires the shared key credential. Three blob types serve different purposes: block blobs (the default, for general files), append blobs (write-once, append-only, suited to logs and audit trails), and page blobs (fixed-size, 512-byte-aligned random read/write, used for VHDs).

Use it for object storage in Azure - file uploads/downloads, static assets, logs (append blobs), or VHD-style random access (page blobs) - in either Node.js or browser code. Several capabilities are Node.js-only (StorageSharedKeyCredential, uploadFile/uploadStream, downloadToFile/downloadToBuffer, SAS token generation, DefaultAzureCredential); browser code is limited to uploadData() and anonymous/SAS-token access, so plan browser uploads around a server-issued SAS token rather than an account key.

Input is blob/container names, data (string, Buffer, stream, File, or ArrayBuffer), and operation options (permissions, expiry, content type). Output is upload/download results, listed blob/container items, blob properties, or a generated SAS token/URL. Install it with:

npm install @azure/storage-blob @azure/identity

Its own best practices: prefer DefaultAzureCredential over connection strings or account keys, use streaming (uploadStream/downloadToFile) for files over 256MB, set correct content types via setHTTPHeaders, issue short-lived SAS tokens for browser client access rather than exposing account keys, check RestError.statusCode for specific error handling (404 not found, 409 already exists, 403 access denied), and use the *IfNotExists methods for idempotent creation.

It's the @azure/storage-blob package paired with @azure/identity for DefaultAzureCredential authentication, generating time-limited SAS tokens for delegated browser or third-party access.

TypeScript/JavaScript developers building file storage into Node.js or browser applications - uploads, downloads, listing, metadata, and delegated SAS-based access - across block, append, and page blob types.