dsync

dsync is a self-hosted tool for one-way data migration and ongoing replication. It is best suited to engineering and DevOps teams that want control over where a transfer runs and need to verify or resume it. The main drawback is its inability to carry source DDL or indexes to the destination.

Overview

dsync copies data between databases and data stores, either as a one-time migration or as an initial copy followed by change capture. That makes it useful when a destination must stay current during a move, but not when both ends need to exchange changes. Operators can run its open source single binary on a laptop, VM, or Docker container, keeping deployment on their own infrastructure.

Progress is reported in a CLI and a web page hosted on the machine running dsync. The ability to resume from saved state helps with interrupted initial copies and change-capture flows. Verification ranges from quick namespace counts to full hash-based checks, so teams can choose between a fast sanity check and a more thorough integrity comparison.

Key features

Connectors and transformations

Supported systems include MongoDB, Cosmos DB through MongoDB API or NoSQL, Azure Document DB, DynamoDB, HBase, AWS DocumentDB, AstraDB, Cassandra, CouchDB/IBM Cloudant, Firestore, PostgreSQL, SQL Server, Oracle, DB2, Weaviate, Qdrant, S3 Vector Index, S3 Storage, and files. The broader connector range gives teams options across databases, vector stores, object storage, and files; however, some connectors are in preview or development, so connector status matters before committing to a migration. Transformations are supported through external processes connected over gRPC.

Operational limits and security

dsync does not replicate source DDL statements or indexes. Teams must account for schema and index setup separately, making it a poor fit when migration needs to reproduce the source structure automatically. Adiom says dsync does not store data and supports network encryption. With a metadata store for resumability, the open source edition has a 16 MB cap on the read plan, a constraint to check against the intended workflow.

The Enterprise edition adds horizontal scalability, observability, and transformations, and is positioned for production work from 100 GB to 100 TB and beyond. Adiom also offers commercial support for both editions. That gives teams a path beyond the open source version, though the Enterprise plan's cost is custom pricing.

Pricing

Open Source

Open Source costs nothing, billed Free forever. It is aimed at small projects and testing, with best-effort support; commercial support is an add-on. The single-binary deployment and core migration, change-capture, verification, and resumability capabilities make it a credible starting point, but the 16 MB read-plan cap with a metadata store and lack of DDL and index replication limit its suitability for some production workflows.

Enterprise

Enterprise uses custom pricing, billed as an annual or monthly fee. It includes commercial support, roadmap prioritization, customization, horizontal scalability, and observability. It is the option for teams with larger production workloads or needs beyond the open source edition, but buyers should weigh its custom-priced terms against their scale and operational requirements.

Platforms

dsync supports Linux, macOS, Windows, web, and self-hosted deployment. The open source binary can run on a laptop, VM, or Docker container, while its web page is hosted on the machine running dsync. Self-hosting suits teams that need to operate the replication process within their own environment; it is not a hosted service.

Who it's for

dsync is a good fit for engineers and DevOps teams at enterprises and fast-growing companies that need one-way migration or replication, value resumability and integrity checks, and can manage deployment themselves. It is less appropriate for teams that require bidirectional sync or expect indexes and DDL to move with the data. Smaller projects and testing have a free path; larger production requirements may call for Enterprise.

Pros and cons

• Resumable transfers: saved-state recovery can avoid restarting an interrupted initial copy or change-capture flow from scratch.
• Choice of verification depth: namespace counts provide a quick check, while hash-based verification offers a more complete integrity comparison.
• Self-hosted operation: the single binary can run on a laptop, VM, or Docker container, giving teams control over where it runs.
• Broad connector coverage: support spans databases, vector stores, S3, and files, though preview and development connectors need checking.
• No schema or index replication: teams must handle DDL and index setup separately at the destination.
• Read-plan ceiling: the 16 MB cap applies to the open source edition when a metadata store is used for resumability.

Alternatives

Consider Database Replication Software for a broader comparison of replication tools.

• SymmetricDS is an alternative with a free Community plan and support for Android and iOS as well as desktop and web platforms; choose it when those platforms are relevant.
• DBConvert Streams offers a free plan centered on an IDE, browser viewers, editor extensions, and federated queries over two sources; choose it when those tools and queries fit better than dsync's migration focus.
• Debezium is a free, open-source distributed change data capture platform under Apache License 2.0; choose it when distributed CDC is the priority.
• NineData offers a free Community plan with on-premise deployment, up to 10 data sources, and up to 10 replication tasks; choose it when those caps suit the workload.
• Artie has a free plan capped at 1M billable rows per month and two active pipelines, with backfills included; choose it when that pipeline and row allowance fits.
• DBSync offers a free Developer plan capped at two connectors, 10,000 records, and three objects; choose it for a small transfer within those limits.
• Google Cloud Datastream has a free tier of 100 GiB of CDC data per month per billing account for AlloyDB and Spanner sources to BigQuery; choose it when that source-destination path fits.
• Oracle GoldenGate has a free perpetual license with a recipe-driven interface and Docker deployment; consider it as an alternative.

Verdict

Choose dsync if you need self-hosted, one-way migration or replication with resumability and verification, and can manage schema and indexes separately. Look elsewhere if bidirectional synchronization or automatic DDL and index replication is essential, or if the open source read-plan cap conflicts with your workflow.