"Plano is an AI-native data plane for AI agents - a framework-friendly, protocol-native fabric that lets you focus on what really matters: your agents' product logic.",
"Plano takes over the plumbing work that slows teams down when handling and processing prompts, including detecting and blocking jailbreaks, routing tasks to the right model or agent for better accuracy, applying context engineering hooks, and centralizing observability across agentic interactions.",
],
diagram:"/IntroDiagram.svg",
},
{
id: 2,
category:"OPEN SOURCE",
title:"",
description:[
"No lock-in. No black boxes. Just an open, intelligent fabric for building more reliable agentic AI applications.",
"Built by engineers with roots in the Envoy ecosystem, Plano brings production-grade reliability to agent traffic and prompt orchestration—while staying fully extensible. Shape it, extend it, and integrate it into your existing workflows without being forced into a rigid framework or a single provider.",
],
diagram:"/OpenSource.svg",
},
{
id: 3,
category:"BUILT ON ENVOY",
title:"",
description:[
"Plano is built on Envoy and runs as a self-contained sidecar alongside your application servers. It extends Envoy's HTTP connection management, filtering, and telemetry specifically for prompt and LLM traffic—so you get production-grade routing, policy enforcement, and observability out of the box.",
"Use Plano with any application language or framework, and connect it to any LLM provider.",
],
diagram:"/BuiltOnEnvoy.svg",
},
{
id: 4,
category:"PURPOSE-BUILT",
title:"",
description:[
"Unlike generic API gateways, Plano is purpose-built for agent workloads, where prompts are the unit of work.",
"Plano treats prompts as first-class traffic: it understands prompt/response flows, tool calls, model selection, and multi-agent handoffs. That means routing, policy enforcement, and observability are optimized for agent execution—not retrofitted from traditional API infrastructure—so your AI applications stay fast, reliable, and easy to evolve.",
],
diagram:"/PurposeBuilt.svg",
},
{
id: 5,
category:"PROGRAMMABLE ARCHITECTURE",
title:"",
description:[
'As agent workloads move beyond prototypes, teams end up scattering critical logic across apps: compliance checks, context "patches," provider-specific quirks, etc. That glue code gets duplicated across agents, is hard to audit, and slows iteration because every policy or workflow change requires touching application code and redeploying.',
"Plano keeps that logic in one place with a programmable Agent Filter Chain—hooks that can inspect, mutate, or terminate prompt traffic early, turning common steps (policy enforcement, jailbreak checks, context engineering, tool gating, routing hints) into reusable building blocks.",