Welcome to Xilos

Xilos is an intelligent agentic AI infrastructure platform by Mill Pond Research (opens in a new tab) that observes, secures, and orchestrates agentic AI across enterprise networks. It operates as a middleware gateway that intercepts, analyzes, and processes every prompt before it reaches an LLM destination.

Info: New to Xilos? Start with the Quickstart — connect your first agent in under 5 minutes. If it speaks the OpenAI API, it works with Xilos.

About Mill Pond Research

Mill Pond Research is an AI infrastructure company building the Unified Control Plane for agentic AI. The company combines two products into a single operational and governance layer:

• WorkBench — Prototype, test, and deploy agents faster with unified access to major AI models, specialized agents, and customizable tools.
• Xilos — Observe, secure, and orchestrate autonomous AI systems with real-time governance, intent-based routing, and a private knowledge base.

Together, WorkBench and Xilos help organizations deploy, secure, orchestrate, and control their entire agentic AI ecosystem — maximizing efficiency and innovation while retaining the visibility and control needed to run AI securely at enterprise scale.

Visit millpondresearch.com → (opens in a new tab) · Visit xilos.ai → (opens in a new tab)

The Three Pillars

Xilos is built on three foundational pillars:

• Observe — Total visibility into all AI traffic. Every query, every model, every dollar — tracked in real time.
• Secure — Local guardrails, PII detection, prompt injection defense, and policy enforcement — all running before any model is called.
• Orchestrate — Semantic caching, multi-model workflows, context compression, and a context engine that compounds knowledge.

Who Is This Documentation For?

Xilos documentation serves three distinct audiences:

• CIOs and CISOs — Strategic, risk-aware stakeholders who need governance, compliance, and security information. Start with Governance and Compliance and Deployment.
• Innovation Leaders — Visionary decision-makers who need speed-to-value and business outcome content. Start with Getting Started and Platform Overview.
• Technical Practitioners — Engineers and architects who need API references, integration specs, and deployment procedures. Start with the API Reference and Guides.

Key Differentiators

• OpenAI-compatible — If your tool speaks the OpenAI API, it works with Xilos. One base URL change.
• Semantic caching — Reuses responses for semantically similar queries. Reduces cost and latency by 50-90%.
• Context compression — Reduces token count by 50-90% before queries reach the LLM.
• Intelligent routing — Route queries to the right model based on cost, latency, or capability. Define rules in natural language — no code required.
• Agent guardrails — Local PII detection, jailbreak prevention, and content filters. No data leaves your infrastructure.
• Context Engine — Every answered query becomes a searchable knowledge asset. Your agents' collective experience compounds.
• Three deployment options — Managed SaaS, Private Cloud (VPC), and On-Premise (air-gapped capable).
• Audit and compliance — Complete immutable trail of every AI interaction. Built for compliance and forensics.

Get Started

Ready to connect your agents? Head to the Quickstart guide, or explore the sections below:

• Getting Started — Quickstart, concepts, and connecting your first agent.
• Platform — Observe, Secure, and Orchestrate pillars.
• Integrations — Connect Hermes, Claude Code, LangChain, and more.
• API Reference — Complete REST API documentation.
• Guides — Step-by-step tutorials for every feature.
• Deployment — SaaS, Private Cloud, and On-Premise options.

Pricing

Xilos offers multiple tiers, from a free plan (no credit card required) to enterprise and on-premise options. View pricing → (opens in a new tab)