# Secure AI Agents

## Build secure, governed AI agents

Provision short-lived, least-privilege sandboxes for every agent. Federate the right context and enforce policy at query time.

[Get a demo](https://meetings.hubspot.com/lukekim/talk-to-sales?uuid=836fd7be-a95e-4cee-b0cb-044fd8ea52a4&utm_campaign=22505145-Website+Demo+Requests&utm_source=homepageheader&utm_medium=website&utm_content=demo+request) [View the docs](https://spiceai.org/docs/use-cases/ai)

### 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

### Agents need context - without risking your data

Generic agent frameworks over-permission access, leading to sensitive data exposure and limited audit trails. Security teams need least-privilege context, inline policy enforcement, and full traces.

### Security is the foundation, not an afterthought

Spice sandboxes, governs, and accelerates agent context in one runtime.

#### Data-centric isolation

Provision short-lived, scoped datasets per task or session. Apply least-privilege access to only the data an agent needs.

[Explore secure AI sandboxing](/content/feature/secure-ai-sandboxing/index.html)

#### Federated context with security controls

Spice presents a single SQL layer over databases, object storage, APIs, and catalogs so applications operate on federated context rather than isolated pipelines.

[Explore SQL federation and acceleration](/content/platform/sql-federation-acceleration/index.html)

#### Inline policy and guardrails

Apply role-based access controls, redact sensitive fields, and restrict tool access by role, dataset, or environment before the prompt ever reaches a model.

[Explore the MCP server gateway](/content/feature/mcp-server-gateway/index.html)

#### LLM gateway with distributed tracing

Route to hosted or local models through the AI Gateway. Capture prompts, retrieve context, tool calls, and outputs for evaluation, red-teaming, and compliance audits.

[Explore LLM inference in Spice](/content/platform/llm-inference/index.html)

### Why teams trust Spice for secure, production-grade AI agents

Spice unifies federated context, least-privilege sandboxes, policy enforcement, and LLM governance so agents are both useful and secure.

- **Native Sandboxing**  
  Scope datasets per agent/task and auto-expire to limit blast radius.

- **Policy Enforcement**  
  Apply guardrails at query time. Restrict tools and enforce roles dynamically.

- **End-to-End Audit**  
  Trace all prompt inputs, context, and model invocations for every agent.

- **Federated and Governed Context**  
  Retrieve only permitted structured and unstructured data.

- **Operational Performance**  
  Local acceleration delivers ms latency context for agents, even with full controls.

- **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

#### “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

#### “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

### Build more performant and secure AI agents

Guides and examples to learn more about deploying secure, high-performance AI agents with Spice.

[Docs](/content/docs/use-cases/ai/index.html)  
**AI Agent Docs**  
Explore the docs for examples and guidance on building AI agents with Spice.

[Recipe](https://github.com/spiceai/cookbook/blob/trunk/guides/security-analyzer/README.md)  
**Intelligent Security Copilot Cookbook**  
Follow this step-by-step recipe to stream query logs into Spice, apply AI-driven pattern analysis, and surface alerts for suspicious database activity.

[Blog](/content/blog/real-time-hybrid-search-using-rrf/index.html)  
**Real-Time Hybrid Search Using RRF: A Hands-On Guide with Spice**  
Surfacing relevant answers to searches across datasets has historically meant navigating significant tradeoffs.