Hi HN, we’re Luke and Phillip, and we’re building Spice.ai OSS - a lightweight, portable data and AI engine and powered by Apache DataFusion & Ballista for SQL query, hybrid-search, and LLM-inference across disaggregated-storage used by enterprises like Barracuda Networks and Twilio.

With v2.0, you can add a real-time analytics node to your operational database with high-throughput native CDC replication.

We first introduced Spice [1] on HN in 2021 and re-launched it on HN [2] in 2024 re-built from the ground up in Rust. In 2025 we launched Spice Cayenne [5], our acceleration and query engine built on the LF+AI Vortex columnar format.

Spice includes the concept of a Data Accelerator [3], which is a way to materialize data from disparate sources, such as other databases, in embedded databases like SQLite, DuckDB, and Spice Cayenne.

Today we're excited to announce Spice 2.0! Our answer to analytics in the era of AI agents: analytical query on operational data - real-time, sandboxed, without ETL.

Bolt an analytics node onto your operational database, including Postgres, MySQL, MongoDB, DynamoDB and run sub-second analytical queries on data that's single-digit seconds fresh, with zero query load on production.

Agents can't act on yesterday's data. That belief has shaped Spice since we founded it in 2021 and it's why we rebuilt replication from the ground up for 2.0.

High-throughput CDC replication, direct from the WAL, binlog, and oplog. No pipelines, no Debezium, no external streaming layer. Native replication bootstraps a 300M-row table in ~9 minutes — ~170x faster ingest than our 1.x Debezium-based path. Start with one table and be querying in minutes.

Spice Cayenne, our Vortex-based acceleration engine, is GA: 1.5x faster than DuckDB acceleration with 3x less memory, and 26x faster than Spice 1.x on TPC-DS. On CH-BenCHmark (TPC-C transactions + TPC-H analytics, concurrently), PostgreSQL sustained 266,861 new orders per minute while Spice served the entire analytical workload and production never ran a single analytical query.

Petabyte-scale, multi-node distributed query built on Apache Ballista is GA: multi-active schedulers, no single point of failure, and 2.9x faster query from one node to three. No JVM, no ZooKeeper. Just Rust and object storage.

Enterprise-ready: a Kubernetes Operator with blue/green deployments and instant rollback, Cedar policy enforced inside the query engine, OIDC, mTLS, and PII masking — all before data ever reaches an agent.

You can read about it at https://spice.ai/blog/spice-2-0-is-now-available and in the Spice OSS release notes [4].

This is just the first version, and we’d love to get your feedback!

GitHub: https://github.com/spiceai/spiceai

[1] https://news.ycombinator.com/item?id=28448887

[2] https://news.ycombinator.com/item?id=39854584

[3] https://github.com/vortex-data/vortex

[4] https://spiceai.org/releases/v2.0-stable

[5] https://news.ycombinator.com/item?id=46316991

Remote sync, MCP, and an API for your Obsidian vault