Unify real-time, governed data for financial services
Spice is a portable SQL query, search, and AI runtime that you can run in your own environment or on Spice Cloud. Query warehouses, operational databases, and object storage in one SQL statement. Accelerate hot data for millisecond reads, and trace queries and model calls in task history.
Do more with your data
100x
up to 100x faster queries
80%
up to 80% cost savings on data lakehouse spend
2x
increase in data reliability for critical workloads
Data fragmentation slows modern finance
Financial institutions depend on real-time data for trading, fraud prevention, and customer insights, but legacy ETL, disconnected search, and siloed analytics create delays and compliance risk. Financial services need governed, low-latency access across systems, and AI that operates transparently within those guardrails.
Governed data. Real-time performance.
Federate, accelerate, search, and manage data and AI for every workflow.
Federate and accelerate SQL across sources
Query disparate data sources in one SQL statement. Push down and materialize frequently accessed data locally for sub-second reads.
Real-time fraud detection
Query the most recent transactions with low latency to find fraud patterns as they occur. Change data capture (CDC) streams inserts, updates, and deletes into accelerated tables, so analytics run on live data.
Intelligent document and record search
Combine vector search with BM25 full-text search across research, contracts, and records. The rrf() SQL function fuses both rankings into one result set.
AI gateway with tracing
Route requests to hosted or local LLMs through one OpenAI-compatible API. Task history records each model call, tool call, and SQL query. Spice Enterprise adds Cedar policies with row filters and column masks.
Built for regulated, low-latency financial workloads
Spice combines federated access, acceleration, hybrid search, and AI in one runtime for financial apps. Deploy it in your own cloud or data center, or use the managed Spice Cloud.
Enterprise-Grade Performance
Materialize hot data locally for real-time apps or fraud detection.
ETL-Free Syncs
CDC from PostgreSQL, MySQL, MongoDB, DynamoDB, or Debezium keeps accelerated tables current without batch jobs.
Hybrid Search
Vector and BM25 keyword search across research, contracts, and records.
Audit and Observability
Record queries, model calls, and tool calls in task history for end-to-end traces.
MCP Server and Gateway
Give AI models governed tools, such as risk models or fraud APIs exposed as MCP servers. Spice serves its datasets to agents over MCP.
Deployment Flexibility
Run Spice anywhere: as a sidecar, microservice, cluster, or on the managed Spice Cloud Platform.
Deployed in production
Run data-intensive workloads on a high-performance engine trusted by teams building real-time systems at scale.
“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
100x
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
Learn how Barracuda gained 100x faster query responses and reduced costs 50% with Spice
“Partnering with Spice AI has transformed how NRC Health delivers AI-driven insights. By unifying siloed data across systems, we accelerated AI feature development, reducing time-to-market from months to weeks - and sometimes days. With predictable costs and faster innovation, Spice isn't just solving some of our data and AI challenges - it's helping us redefine personalized healthcare.”
Tim Ottersburg
VP of Technology, NRC Health
“Spice AI grounds AI in our actual data, using SQL queries across many data sources. This brings accuracy to probabilistic AI systems, which are very prone to hallucinations.”
Rachel Wong
CTO, Basis Set
Trusted by global enterprises
Build more robust financial applications
Guides and examples for building secure, low-latency financial applications with Spice.
Hybrid Search Docs
Learn how to combine semantic, full-text, and keyword search with Spice's hybrid search capabilities
Federation and Acceleration Docs
Learn how to get started with Spice federation and acceleration. Query, join, and accelerate data using SQL from multiple sources, including databases, data warehouses, and data lakes
Making Object Storage Operational for Real-Time and AI Workloads
TLDR Introduction Although legacy systems and workflows remain common, many enterprises are re-evaluating their architectures to meet new demands – driven in part, but not exclusively, by AI – that require support for more data-intensive and real-time applications. The underlying storage needs for these novel workloads are generally outside the bounds of a traditional operational […]
FAQs
Common questions about Spice for banks, payments, insurance, and capital markets teams
A data and AI platform for financial services gives applications and AI models governed, low-latency access to financial data. Spice does this with one runtime for SQL federation and acceleration, search, and model serving. Teams use it for fraud detection on live transactions, document search, and AI assistants. It queries Snowflake, Databricks, PostgreSQL, Amazon S3, and other supported sources in place, with no new ETL pipeline.
Yes. Spice accelerates transaction tables in memory or on local disk for millisecond reads. Set refresh_mode: changes to use change data capture from PostgreSQL, MySQL, MongoDB, DynamoDB Streams, or Debezium. Fraud queries then run on current rows instead of a nightly extract.
Not when you self-host Spice in your own cloud account, Kubernetes cluster, or data center. Federated queries read the source in place, and accelerated copies stay in your runtime. A hosted model receives each prompt and the rows that Spice adds as context, so use a local model to keep that data in place. With the managed Spice Cloud, the runtime and its accelerated data run on Spice infrastructure.
The open-source runtime requires no auth until you configure API keys or mutual TLS (mTLS) in spicepod.yaml. API keys without a :rw suffix and datasets without access: read_write are read-only. Spice Enterprise adds OIDC sign-in (in preview) and Cedar policies with row filters, column masks, and roles. Task history, on by default, logs each query, model call, and tool call to spice.runtime.task_history for 8 hours, unless you set runtime.task_history.retention_period.
Yes, Spice connects to MCP servers, such as a fraud API, and the tools parameter limits which tools each model can call. Spice also serves its datasets to external agents at /v1/mcp, which requires API key auth. By default the runtime listens on 127.0.0.1:8090 and accepts only a localhost, 127.0.0.1, or ::1 Host header. For remote agents, set --http to a reachable address, list hosts in runtime.mcp.allowed_hosts, and use TLS or a private network.
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 (opens in a new tab)