- Products
- Solutions
Real-Time Analytics
Real-time database for real-time data warehousing, customer-facing and agent-facing analytics
Lakehouse Analytics
The fastest lakehouse SQL engine, replacing Trino/Presto, SparkSQL
Observability in the AI Era
The most cost-effective alternative to Elasticsearch observability
VeloDB for AI
The AI-Ready analytics database for the AI era
- Docs
- Resources
- Pricing
- Contact us
High-performance CDC into VeloDB — direct or via Iceberg.
Supermetal is a high-performance CDC engine for operational databases — Postgres, MySQL, Oracle, SQL Server, and MongoDB. It ships as a single Rust binary built on Apache Arrow, so changes flow into VeloDB in seconds. Pick the path that fits: write directly into VeloDB, or land first in Apache Iceberg and let VeloDB query the lakehouse — no Kafka cluster, Flink job, or Spark pipeline in between.
What Supermetal does
Supermetal runs as a self-contained agent with no external dependencies. Inside the binary, Apache Arrow serves as the in-memory format — preserving precision and types end-to-end and avoiding the JSON / IPC overhead that plagues conventional pipelines. The result is a streaming, single-process replicator that can move terabytes on modest hardware (2 vCPU, 8 GB RAM) while preserving transactional consistency from source to target.
Why VeloDB + Supermetal
Pick the path to VeloDB
Replicate operational data straight into VeloDB tables, or land it in Apache Iceberg first and query the lakehouse with VeloDB. Same agent, same source DBs — choose what fits your stack.
Iceberg sink without Flink or Spark
When you take the Iceberg path, Supermetal writes directly from the database log — no intermediate streaming compute to stand up, tune, or pay for. VeloDB then reads the same Iceberg tables.
Database CDC, not SaaS connectors
Purpose-built for operational databases: WAL/oplog-based capture from Postgres, MySQL, Oracle, SQL Server, and MongoDB, with parallel snapshots that finish in a fraction of the time conventional tools take.
Single Rust binary, Apache Arrow inside
No Kafka, no JVM, no orchestration sidecars. Rich types are preserved end-to-end via Arrow, with zero-copy in-process movement — same low footprint on EC2, Kubernetes, or a hardened on-prem host.
What you can build
Real-time analytics on operational data
Stream changes from your transactional databases straight into VeloDB and serve sub-second analytics without touching the primary.
Lakehouse without the streaming compute
Land Postgres / MySQL / Oracle / SQL Server / MongoDB changes directly in Iceberg — skip the Kafka + Flink/Spark pipeline and let VeloDB query the lakehouse.
On-prem and air-gapped CDC
A single Rust binary deploys cleanly into restricted environments. Pair with a self-hosted or BYOC VeloDB to keep data inside your trust boundary.
How the data flows
Ready to get started?
Tell us about your stack and use case. Our team will help you stand up VeloDB with Supermetal for your environment.
