SUPERWISE has launched Sentinel, an enterprise AI gateway that redacts secrets, personally identifiable information (PII), and toxic content in real time. The solution is designed to give organizations a live guardrail layer over every large language model (LLM) interaction.
TL;DR
- SUPERWISE introduces Sentinel, an enterprise AI gateway that redacts sensitive data in real time.
- Sentinel is deployed at the network edge or locally on user's PCs, ensuring raw data never leaves the customer's perimeter.
- The solution is designed to protect against the risks defined in the OWASP LLM Top 10, including prompt injection, data leakage, jailbreaks, and unsafe output.
What happened
SUPERWISE, the Agentic Management Platform, has announced the launch of Sentinel, an enterprise AI gateway. The solution is designed to give organizations a live guardrail layer over every large language model (LLM) interaction in their business. Sentinel is deployed as a lightweight gateway at the network edge or locally on user's PCs, protecting company and customer data in AI traffic by redacting secrets, personally identifiable information (PII), and toxic content in real time. According to SUPERWISE, the solution ensures raw data never leaves the customer's perimeter.
Sentinel is built to protect against the risks defined in the OWASP LLM Top 10, including prompt injection, data leakage, jailbreaks, and unsafe output. It works across any model or provider, including OpenAI, Claude, and Gemini. The solution deploys via Docker, the CLI, or directly into a customer's container cluster, without requiring security code to be added to any individual application.
Why it matters
As AI adoption accelerates across engineering and business teams, most organizations have little to no visibility into how employees and applications are actually using large language models. Sentinel addresses that blind spot directly, giving security and compliance teams a single control plane to answer the question of what AI is allowed to say and whether the organization can prove it.
For developers and startups, Sentinel offers a way to integrate AI into their workflows without compromising on security and compliance. The solution's ability to work across any model or provider makes it a versatile tool for organizations using multiple AI services. For investors, SUPERWISE's focus on enterprise AI governance and security positions it as a key player in the rapidly growing AI market.
Key facts
- Sentinel is designed to redact secrets, PII, and toxic content in real time.
- The solution is deployed at the network edge or locally on user's PCs.
- Sentinel works across any model or provider, including OpenAI, Claude, and Gemini.
- The solution deploys via Docker, the CLI, or directly into a customer's container cluster.
- Sentinel is built to protect against the risks defined in the OWASP LLM Top 10.
- The solution follows a three-step model: deploy a sentinel proxy at the edge, route AI traffic through it, and capture, redact, and log actions.
- Sentinel is available now and can be installed via the CLI (Python 3.11+) and deployed in minutes.
- SUPERWISE works with organizations across industries including healthcare and manufacturing.
Context
The launch of Sentinel comes at a time when AI adoption is accelerating across various industries. As organizations increasingly integrate AI into their workflows, the need for robust governance, security, and compliance solutions has become paramount. SUPERWISE's Agentic Management Platform is designed to help enterprises govern, secure, and scale their use of AI. The platform includes Sentinel, along with other capabilities such as Chat, Agent Studio, Guardrails, and Observability.
The OWASP LLM Top 10, which Sentinel is designed to protect against, outlines the top security risks associated with large language models. These risks include prompt injection, data leakage, jailbreaks, and unsafe output. By addressing these risks, Sentinel aims to provide organizations with a comprehensive solution for securing their AI traffic.