---
id: blobfile
version: "3.2.0"
license: Public Domain
license_treatment: permissive
maintenance: active
---

# blobfile — Read GCS, ABS and local paths with the same interface, clone of tensorflow.io.gfile

License: permissive · Maintenance: active · Downloads: 9.6M/mo

## What it is and what it does

Blobfile abstracts away the differences between local filesystem operations and cloud blob storage by exposing a single Python interface modeled on the built-in `open()` function. You can read and write files on Google Cloud Storage (gs://), Azure Blob Storage (az:// or https://), and local paths using the same code, with optional streaming to avoid downloading entire files into memory.

The library provides both high-level file operations (BlobFile for opening, copy for remote-to-remote transfers, exists, listdir, walk, stat, remove, rmtree) and path utilities (basename, dirname, join) that work uniformly across all three backends. It handles authentication through environment variables or application default credentials, and offers configuration options for connection pooling, timeouts, chunk sizes, and caching behavior.

Use it for:

- Process large datasets stored in GCS or Azure without downloading them entirely to local disk first
- Write data processing pipelines that work with both local development files and cloud-hosted production data
- Migrate file operations between cloud providers by changing only the path prefix, not the code
- Implement parallel blob processing with built-in support for concurrent operations across cloud storage
- Cache remote files locally for repeated reads while maintaining a unified file interface

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Provides a unified Python file interface for reading and writing to local filesystems, Google Cloud Storage, and Azure Blob Storage using an API similar to the built-in `open()` function.

Yes, if you work with cloud blob storage and want a simpler, provider-agnostic file interface. The low install friction, active maintenance, public domain license, and lack of known vulnerabilities make it a safe choice. Install only if you actually need multi-cloud or local-plus-cloud file abstraction; if you use only one provider, that provider's native SDK may be more direct.

## Install

pip install blobfile
uv add blobfile
poetry add blobfile

## Installing blobfile

Before you install: Low friction: pure Python wheel with four straightforward runtime dependencies (pycryptodomex, urllib3, lxml, filelock). Repository is active with recent commits and no known vulnerabilities.

License in practice: Public Domain license means no restrictions on use, modification, or distribution—you can use this freely in any project without license compatibility concerns.

Quickstart:

pip install blobfile

import blobfile as bf

with bf.BlobFile("gs://my-bucket-name/cats", "wb") as f:
    f.write(b"meow!")

with bf.BlobFile("gs://my-bucket-name/cats", "rb") as f:
    print(f.read())

Requires Python 3.10 or later. Authentication to GCS or Azure requires appropriate credentials (environment variables or default application credentials).

Verify before relying:

- Performance characteristics and latency overhead compared to cloud provider SDKs
- Concurrent access limits or connection pooling behavior under high load
- Compatibility with specific GCS/Azure API versions or deprecated endpoints

## Package facts

- License: Public Domain (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 9.6M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

cloud storage file interface, gcs azure blob unified api, remote file access python, google cloud storage client, azure blob storage reader, streaming file operations, multi-cloud file abstraction, cloud-storage, file-abstraction, multi-cloud

[View on SkillFed](https://skillfed.io/packages/blobfile) · [View on PyPI](https://pypi.org/project/blobfile/)