Ergonomic for a single backend
Read, write, stat, list, copy, and delete through one Operator, even when your application only uses S3 or a local directory. Paths, metadata, and error kinds follow one model across operations.
Apache OpenDAL™ — Open Data Access LayerOne Layer,
One Layer,
All Storage.
Write clean storage code from simple reads and writes to streaming, byte-range reads, and concurrent uploads — on a single S3 bucket or across 50+ backends.
50+services
17languages
5k+stars
10M+downloads
$ cargo add opendal
use opendal::services::S3;
use opendal::Operator;
// Configure a backend once, then use one operator.
let builder = S3::default().bucket("data");
let op = Operator::new(builder)?;
op.write("hello.txt", "Hello, World!").await?;
let bytes = op.read("hello.txt").await?;
Why OpenDAL
Clean application code, from the first file to heavy I/O.
Adopt OpenDAL for one bucket or local directory today, then add large transfers, resilience policies, and more backends while your storage code keeps the same shape.
Built for demanding I/O
Stream large objects, read only the byte ranges you need, and upload parts concurrently on services with multipart or block writes. Add retries, timeouts, and concurrency limits as layers.
Portable across 50+ services
Use memory in tests, the local file system in development, and S3, GCS, Azure Blob, HDFS, or a database in production. Capability checks show where services differ.
Capabilities
From simple reads to heavy I/O.
One operator covers byte-range reads, concurrent transfers, file operations, and production policies. Examples use the Rust core; each language guide shows the options and layers its binding exposes.
let op = Operator::new(S3::default().bucket("data"))?;
// Read just the bytes you need.
let head = op.read_with("logs/today").range(0..64 * 1024).await?;
// Or read a large object in concurrent chunks.
let object = op
.read_with("big.parquet")
.concurrent(8)
.chunk(8 * 1024 * 1024)
.await?;
Used by
Case studyHow ScopeDB accesses object storage with OpenDALA columnar database on object storage uses one operator for sparse range reads, concurrent multipart writes, cloud authentication, and separate timeout and retry policies for reads and writes.Powering AI, analytics, and real-time data
OpenDAL runs in production across the open-source ecosystem. These are some of the projects that build on it.
Services
50+ storage services, one interface.
Enable only the backends your application needs. Core operations share one interface, and each service reports optional features through its capabilities.
s3
gcs
azblob
oss
cos
obs
b2
tos
fs
hdfs
alluxio
goosefs
lakefs
ipfs
dbfs
gdrive
dropbox
onedrive
hf
github
koofr
http
ftp
webdav
sftp
sqlite
mysql
postgresql
mongodb
surrealdb
d1
redis
etcd
rocksdb
memcached
tikv
foundationdbStart building
One layer for all your storage.
Start with one bucket or directory, and keep the same operator model as your workloads and backends grow.