Real-time analytics on operational data, without ETL
Add a sandboxed, analytics-ready replica alongside PostgreSQL, MySQL, and MongoDB in minutes. Spice keeps it fresh with high-throughput change data capture for sub-second queries, no ETL, and no analytical load on your production database.
Real-time analytics, zero production impact
0
analytical queries run on your production database, ever
0s
end-to-end freshness under load, from source commit to query-ready
0x
faster than DuckDB with 3x less memory on Spice Cayenne
Operational databases weren't built for analytics
Running analytical queries directly on MySQL, PostgreSQL, or MongoDB puts mission-critical operations at risk: heavy queries contend with production traffic, row-level security policies grow brittle, and data can leak beyond its intended audience. The usual workaround, ETL pipelines that copy data into a separate analytical system, is expensive to build, costly to operate, and never real-time, with data arriving hours or days stale.
Add analytics to your operational data, without ETL
Point Spice at a live operational database and it maintains a continuously updated analytics replica. Start with one table, then compose and join across replicated sources in a single SQL query.
Zero load on production
Spice replicates from the native change log, so analytical queries and AI agents run against the sandboxed replica. Your operational database never executes a single analytical query.
Explore CDC replication
Real-time freshness, no pipelines
Native replication from the PostgreSQL WAL, MySQL binlog, and MongoDB oplog delivers single-digit-second freshness. No Debezium or external streaming layer required.
See real-time CDC
Sub-second query performance
Spice materializes the replica with the Spice Cayenne accelerator, built on the Vortex columnar format: 1.5x faster than DuckDB with 3x less memory. DuckDB and SQLite acceleration engines are also supported.
Explore Spice Cayenne
Connect the tools you already use
Query the replica from Microsoft Power BI, Tableau, Looker, and Apache Superset over Arrow Flight SQL, ODBC, and JDBC, or build in code with the SpicePy Python library and the Go, Rust, Java, and JavaScript SDKs.
Read the Power BI guide
Why teams run analytics on Spice
Spice adds an analytics replica to your operational data with the freshness, scale, and controls that production analytics and AI agents demand.
Native CDC replication
Replicate directly from the PostgreSQL WAL, MySQL binlog, and MongoDB oplog, with automatic slot management and bootstrapped snapshots. No Debezium required.
One SQL surface
Federate and join replicated datasets with object storage, warehouses, and APIs in a single query using standard SQL.
Sub-second acceleration
Materialize hot working sets locally with Spice Cayenne, built on the Vortex columnar format, for millisecond query performance.
Petabyte-scale distributed query
Scale beyond one node with multi-node distributed compute built on Apache Ballista, object-store native and highly available.
Sandboxed and governed
Each replica is a physically isolated sandbox. Row- and column-level policy, PII masking, mTLS, and OIDC authentication are enforced inside the query engine.
Built on open standards
Arrow, Iceberg, Delta, Parquet, and Vortex, served over Arrow Flight SQL, ODBC, and JDBC. Open source and portable, with no lock-in.
Query operational data with SQL in the Spice portal
Explore datasets, run SQL, and inspect results in the Spice Cloud portal. Built-in observability traces every query with execution timing and lineage, so you can validate freshness and tune performance against the replica instead of production.
Proven in production
Teams building real-time systems at scale run Spice to serve analytics and AI from operational data.
0x
Faster queries
“It just spins up and works, which is really nice. The responsiveness is amazing, which is a huge gain for the customer.”
Darin Douglass
Principal Software Engineer, Barracuda
“Spice opened the door to take these critical control-plane datasets and move them next to our services in the runtime path.”
Peter Janovsky
Software Architect, Twilio
FAQs
Learn how an analytics replica keeps operational data query-ready without ETL or load on production.
What is an analytics replica?
An analytics replica is a read-optimized, continuously updated copy of your operational data that Spice maintains alongside your database. Analytical queries, dashboards, and AI agents run against the replica instead of production, so heavy reads never contend with your transactional workload.
Does it add load to my production database?
No. Spice replicates from the native change log, the PostgreSQL WAL, MySQL binlog, or MongoDB oplog, rather than polling with queries. In benchmarks the operational database sustained its full transactional workload while Spice served all analytics from the replica, and production never executed a single analytical query.
Which databases can I replicate from?
Spice ships native replication for PostgreSQL, MySQL, MongoDB, and DynamoDB, plus append-only event streams from systems like Kafka with upsert semantics. Debezium is also supported. See the CDC documentation for the full list.
How fresh is the data?
End-to-end freshness is single-digit seconds under load, roughly 2.0 seconds for CDC covering inserts, updates, and deletes, measured from source commit to exactly-correct results on the replica.
Do I need to build ETL pipelines?
No. Replication is built into the runtime, so there are no external ETL jobs, no Debezium to operate, and no streaming layer to manage. It is incrementally adoptable: start with one table and expand as needed.
Which BI tools and clients can query the replica?
Any tool that speaks Arrow Flight SQL, ODBC, or JDBC. That includes Microsoft Power BI through the native Spice connector, along with Tableau, Looker, and Apache Superset. In code, use the SpicePy Python library or the Go, Rust, Java, and JavaScript SDKs.
How does analytics scale beyond a single node?
Spice adds multi-node distributed compute built on Apache Ballista for petabyte-scale query. Distribution composes with local acceleration: use acceleration for working sets that fit on one node, and distribution for data that does not.
Is Spice open source?
Yes. The Spice runtime is open source under the Apache 2.0 license, including federation, acceleration with Cayenne, hybrid search, and AI integration. Spice Cloud offers a fully managed option, and Spice.ai Enterprise adds self-hosted deployment with SSO, RBAC, and audit logs.
Learn more about analytics on Spice
Guides, benchmarks, and references for adding a real-time analytics replica to your operational data.
Spice 2.0: Real-Time Analytical Query on Operational Data, Without ETL
Add real-time analytical query and search to operational data without ETL via high-throughput CDC replication, petabyte-scale distributed compute built on Apache Ballista, and enterprise-grade controls.
Analytics Replica
See how Spice adds a sandboxed analytics replica to PostgreSQL, MySQL, and MongoDB with real-time CDC replication.
Change Data Capture (CDC) Replication
Configure native replication from the PostgreSQL WAL, MySQL binlog, and MongoDB oplog, plus event-stream ingestion from Kafka.
See Spice in action
Walk through your use case with an engineer and see how Spice handles federation, acceleration, and AI integration for production workloads.
Talk to an engineer