LCLangChainAGAG-UIAngular

The Enterprise Agent Framework for Angular

Signal-native chat UI for any agent runtime — LangGraph, AG-UI, or your own backend.

npm install @ngaf/chat @ngaf/langgraph

What makes the Angular Agent Framework different

Runtime

One Angular UI. Any agent runtime.

Same primitives drive LangGraph, AG-UI, CrewAI, Mastra, Pydantic AI, AWS Strands, and your own backend.

Streaming

LangGraph streaming for Angular.

agent() ships LangGraph streaming for interrupts, branch and history, tool progress, and tool results — plus error(), status(), and reload().

Generative UI

Generative UI, built in.

Render Vercel json-render and Google A2UI specs into Angular components. No second framework to bolt on.

License

MIT. Headless primitives, drop-in compositions.

No tier gates on Angular. Use the unstyled primitives, or the opinionated chat shell — your call.

The Last Mile Problem

Most AI projects get close.
Almost none ship.

The issue is not generating a demo. It is shipping a trustworthy product.

66%

of AI solutions are almost right — not quite production-ready

50%

of GenAI projects abandoned after proof of concept

75%

of developers still want a human in the loop when trust breaks down

Your backend agent may already work. The frontend and production path is what slips the schedule.

The Solution

From proof of concept to production in 90 days

The structured pilot engagement that closes the last-mile gap.

1

Pilot

We assess your LangGraph agent and Angular codebase. You get a concrete production plan.

2

Build

We work alongside your team to integrate the framework. Streaming, generative UI, testing — done right.

3

Ship

Your agent goes to production with your team owning every line. No vendor lock-in, no black boxes.

Explore the Pilot Program →

Included with every app deployment license

The Cacheplane Stack

Four libraries. One architecture.
Every layer you need.

Whatever your agent runtime — LangGraph, AG-UI, or your own backend — these libraries ship it.

Agent

@ngaf/langgraph

The reactive bridge to LangGraph

Signal-native streaming connects your Angular templates directly to LangGraph agent state. Interrupts, persistence, time-travel, and branch history — every LangGraph feature has a first-class Angular API. Test deterministically with MockAgentTransport.

Angular SignalsLangGraph CloudMockAgentTransport
Explore Agent →
Gen UI

@ngaf/render

Agents that render UI — on open standards

Built on Vercel’s json-render spec and Google’s A2UI protocol. Your agent emits a render spec, your Angular components render it — with streaming JSON patches, component registries, and signal-native state. No coupling between agent logic and frontend.

Vercel json-renderGoogle A2UIJSON Patch streaming
Explore Render →
Chat

@ngaf/chat

Production chat UI in days, not sprints

Every component you need — streaming messages, tool call cards, interrupt panels, generative UI, debug overlay, theming — pre-built, composable, and WCAG accessible. Built on the Agent and Render libraries so everything works together from day one.

<chat-messages><chat-debug>WCAG accessible
Explore Chat →
AG-UI

@ngaf/ag-ui

Run any AG-UI compatible backend

AG-UI is the open agent-to-UI protocol used by CrewAI, Mastra, Microsoft Agent Framework, AG2, Pydantic AI, AWS Strands, and the CopilotKit runtime. The @ngaf/ag-ui adapter wraps an AbstractAgent into the same Agent contract @ngaf/chat consumes — same UI, different runtime.

CrewAIMastraMicrosoft AFCopilotKit
Explore AG-UI →

Get the Guide

or download directly

What's Inside

From Prototype
to Production

The Angular Agent Readiness Guide. Six chapters. Six production-readiness dimensions. What separates demos from shipped products.

Ready when you are

Half of GenAI projects die after proof of concept. Angular Agent Framework closes the gap. Start with a conversation.

Start Your Pilot →Download the Guide

App deployment license · $20,000 · 3-month co-pilot engagement