The workbench that moves data anywhere
Bridge is a self-hostable data integration workbench that moves your data between any systems — safely and simply. Connect 100+ heterogeneous sources and targets, capture real-time changes non-intrusively from database logs, shape and clean data on a visual low-code canvas, and run elastic, high-throughput pipelines across databases, data lakes, files, and streaming — without writing custom code.
What's inside
Log-based CDC
Capture inserts/updates/deletes by parsing source redo/binlog/WAL logs. Capture runs as a lightweight process against your database logs.
100+ connectors
Relational, NoSQL, data lakes, warehouses, files, message queues, IoT and time-series databases.
Visual + IaC
Drag-and-drop canvas for analysts; declarative config + GitOps for platform teams. Same pipeline, two modes.
Transform in-flight
Filter, validate, mask sensitive fields, lookup, route, and script (JS/Java/SQL) without a separate dbt step.
Schema auto-mapping
Heterogeneous DDL inference and type mapping. Library-level sync with automatic table creation.
Run cockpit
Live throughput, CDC lag (P95), and per-pipeline spend — no more guessing what a sync costs.
100+ connectors supported today
The complete list, by category.
Databases
Oracle, MySQL, SQL Server, PostgreSQL, Db2, MongoDB, Redis, Dameng, KingbaseES, GBase
Data warehouses
Snowflake, Redshift, BigQuery, ClickHouse, Greenplum, AnalyticDB
Data lakes & lakehouses
S3, HDFS, Hive, Iceberg, Delta, MaxCompute, HUAWEI FusionInsight
Files
FTP/SFTP/SMB, OBS/OSS/Ceph, CSV/XML/JSON/Excel/Parquet
Streaming & MQ
Kafka, RocketMQ, RabbitMQ, Pulsar
IoT & time-series
MQTT, InfluxDB, TDengine, Druid
Missing a connector? Talk to our team — we'll help you get connected.
Point at a source. We read its logs — never its tables.
Bridge captures inserts, updates, and deletes by parsing redo / binlog / WAL logs directly. Capture reads the transaction log with negligible overhead on your production database — and the initial bulk load plus continuous CDC live in one job.
- Oracle, MySQL, PostgreSQL, SQL Server, Db2, MongoDB — capture reads the source transaction logs
- Initial bulk load + continuous CDC in a single pipeline
- Automatic DDL capture keeps schema drift from breaking the sync
Shape the data before it lands
Filter, validate, mask sensitive fields, enrich via lookup, and route — inside the same job, with no separate dbt step. A low-code canvas for analysts, a code mode for platform teams, and the same pipeline either way.
- Low-code visual mapping plus SQL / JS code mode for GitOps
- Schema-drift handling and automatic type mapping
- Reusable transforms shared across every pipeline
source.postgres.orders| where status = 'paid'| mask(pii_email, sha256)| lookup(currency_rates, fx)→ snowflake.analytics.orders_clean
A run cockpit you can trust
Live throughput, CDC lag at P95, and per-pipeline spend — surfaced in real time, with the 3× monthly spend cap enforced by the engine itself. No more guessing what a sync actually costs.
- Per-pipeline cost, not per-connector billing
- CDC lag and error alerts on every run
- Audit trail for compliance and rollback
Install anywhere, activate with a key
A single license key unlocks the full platform. Pull the image, point it at your source and target, and sync.
$ docker pull sumetaflow/engine:latest# or: brew install sumetaflow (macOS / Linux)$ sumeta activate --key <YOUR_LICENSE_KEY>$ sumeta pipeline create --from postgres --to snowflake✓ CDC capture active · log-based capture · spend cap 3× enforced
Self-host by default
Your data never leaves your account. Run on a laptop, a single VM, or a multi-node cluster.
On-prem / isolated
For regulated and isolated networks, sync via file exchange across security gateways. No inbound connection required.
License, not lock-in
A signed license key activates the software. Cancel anytime — your pipelines and data remain yours.
See it move your data
Download a free evaluation license, or book a 30-minute architecture walkthrough.