AG-UI

Easily bring AI agents into user-facing applications

[Avi Chawla @_avichawla]

Anthropic's in trouble, again!

They spent years building what's now fully open-source.

What made Claude feel different from a normal app is that the agent could act inside the interface instead of only talking in a chat box.

For instance, Claude Artifacts let an agent render real UI, charts, dashboards, and interactive components that assemble live inside the response.

Every major AI product tried to replicate it.

But the problem was that unlike reasoning, planning, tool-calling, etc., none of it shipped natively with LangGraph, CrewAI, or Google ADK.

So teams started building an owned version that required engineering the entire interface layer from scratch.

Most teams, however, just settled for shipping the agent as a backend API in a chat box since rendering the UI is only one piece of it.

To actually make it work, the interface layer also needed real-time streaming, state kept in sync between agent and UI, conversations that persist across sessions, and reconnection when a user refreshes mid-run.

@CopilotKit is now the only open-source framework that actually lets you build your own full-stack Claude-like apps.

It decouples the agent from the interface, talking over AG-UI (an open protocol for agent-to-user communication).

Being a standard protocol, the frontend never needs to know whether it is talking to a LangGraph or a CrewAI agent. You can change the backend anytime and the UI will never notice.

In practice, CopilotKit's interface layer gives several pre-implemented React building blocks that wire the agent directly into the app, like:

- generative UI, so the agent renders real components instead of text
- chat windows, sidebars, and popups, or a fully headless setup
- shared state, so the agent and app stay in sync
- human-in-the-loop approvals, where the agent waits before acting
- persistent threads that store the whole session, including the agent-user interactions and generated UI, not just text

And because that full history is captured, those interactions can feed a self-learning layer that also improves the agent from real usage over time.

The interface layer that Anthropic spent years engineering in-house is now literally available to any developer/team.

CopilotKit is open-source with 30k+ GitHub stars, and AG-UI, the protocol underneath, is already supported across every major agent framework: LangGraph, CrewAI, Mastra, Google ADK, and more.

CopilotKit GitHub repo → https://t.co/wkQ1taF0rM

(don't forget to star it ⭐ )

If you want to go deeper, I found a detailed breakdown by Shubham Saboo recently on the three Generative UI patterns, with implementation.

Read it below.

[CopilotKit🪁 @CopilotKit]

Chat was the first UI for AI. Generative UI is the next one.

CopilotKit lets your agent render real React components,
Call frontend tools, and drive the actual app.

Open source, built on AG-UI.

Start building here 👇
https://t.co/mFTyD9jvjp

[Shubham Saboo @Saboo_Shubham_]

https://t.co/w7v8Vq1JNN

[CopilotKit🪁 @CopilotKit]

Claude Agents just got a UI in ANY app. ✨

Community developer Gage Krumbach (https://t.co/3QUwruxxb1) shipped the AG-UI integration for the Claude Agent SDK.

Until now, a @claudeai Agent lives in your terminal - the users who'd actually use it never see it.

CopilotKit changes that: it uses AG-UI as the transport layer, streaming your Claude Agent SDK agent straight into the frontend as a live copilot people can actually interact with.

[CopilotKit🪁 @CopilotKit]

CopilotKit just went mobile with React Native! 🚀

Build agentic apps on iOS and Android, not just the web.

[Marktechpost AI @Marktechpost]

The gap between a demo agent and a production agent is memory.

CopilotKit just launched the Enterprise Intelligence Platform — a managed persistence layer for agentic applications built on the open-source CopilotKit stack.

The core primitive: Threads — persistent session objects that capture generative UI, human-in-the-loop workflows, shared state, voice, files, and multimodal interactions across sessions and devices.

No custom storage layer. No hand-rolled memory infrastructure. Agents retain context, state, and history out of the box.

What's on the roadmap — — Analytics & Insights — real-time dashboard, SQL-queryable data lakehouse, OTLP export to DataDog and NewRelic — Self-Improvement — Continuous Learning from Human Feedback (CLHF), prompt mutation, per-user adaptation — no fine-tuning required

Read the full technical breakdown: https://t.co/hvgIFYbGw2

@CopilotKit

[Atai Barkai @ataiiam]

We've raised $27M to build @CopilotKit — the Agentic Frontend Stack connecting humans & agents.

Because all UI will be AI.

Co-led by Glilot Capital, NfX and SignalFire.

[Atai Barkai @ataiiam]

🎉 Excited to announce our upcoming course with @DeepLearningAI

Fullstack Agents & Generative UI

All UI will be AI.
Every interaction with technology will be mediated by agents.

In this course we walk through:
✨ The Generative UI Spectrum — controlled, declarative, fully open
🛠️ Building with A2UI, AG-UI and MCP Apps
🧱 Building Blocks for FullStack, Cursor-like Agent Applications

This is the most complete walkthrough of Generative UI you'll find anywhere - the foundation of the next generation of software.

Huge thanks to @AndrewYNg and the DeepLearning AI team. We put a lot of heart into this one and we're really proud of how it turned out.

Join the waitlist:
https://t.co/tRkNuYkGNT

[Atai Barkai @ataiiam]

We open sourced Claude Artifacts for your own apps.

Ask your agent to visualize algorithms, create 3D animations, render charts, or generate interactive diagrams.

PLUS it runs the generated code securely w/ sandbox iframe.

It's additive to your app, so you can drop it in with one toggle.

➡️ https://t.co/oCOZKDatwn

[CopilotKit🪁 @CopilotKit]

Is Sama summoning AG-UI?

[CopilotKit🪁 @CopilotKit]

It's the golden age of building.

Soon, all UI will become AI.

While Agents need one thing, Users need another.

Make your agent speak Generative UI via the three different patterns, in minutes. This was generated by an agent on the fly, in real time 👇

[CopilotKit🪁 @CopilotKit]

https://t.co/xbvukHDXM8

[CopilotKit🪁 @CopilotKit]

Anthropic is going all in on this, and now you can too 👇

Let your agents completely generate interactive UIs on the fly that stream live into the chat.

✨ Introducing Open Generative UI

Enable it by setting openGenerativeUI: true

That's it. Completely open source.

[CopilotKit🪁 @CopilotKit]

@openclaw and @tan_stack are using AIMock 🦞🏖️

1M weekly downloads by tomorrow...

[Nathan Tarbert @nathan_tarbert]

> be a dev building an AI app
> write tests like a responsible person
> every CI run hits real OpenAI, Claude, Pinecone
> every flaky retry burns more tokens
> try to mock it yourself
> mocks drift from the real APIs in a week
> give up, pay the OpenAI tax
> then @CopilotKit drops AIMock
> within two weeks, weekly downloads are at 700K+
> even @openclaw is using in their CI
> one command to mock everything your AI app talks to
> dev finally sleeps at night

AIMock boots 11 LLM providers, MCP, A2A, AG-UI, and vector DBs on a single port.

$0 to run, open source, zero dependencies

Record real APIs once, replay forever. Chaos test with 500s, malformed JSON, and mid-stream drops.

Your tests stop depending on the same providers you're building on.
> https://t.co/IpU6cbsahY

[CopilotKit🪁 @CopilotKit]

AIMock already has 700,000 weekly downloads + TONS of new features for testing.

https://t.co/zBxQ3QaGwR

[CopilotKit🪁 @CopilotKit]

Calling all agent builders 📣

The Generative UI Community Demo Fest is tomorrow.

3 minutes, your code, your peers vote.

Meta Ray-Bans + CopilotKit hoodie for the winner →
It's not too late to register. See you there!
👉https://t.co/uznWSVXHwu

[Manfred Steyer @ManfredSteyer]

🧵 Agentic UI gets messy fast. AG-UI is the protocol that brings order to it.
(👇 full article at the end)

It keeps the frontend decoupled from the agent and the LLM.

Big upside: your frontend stays independent from the server stack!
1/6

[CopilotKit🪁 @CopilotKit]

Testing just an LLM? ✅
You might need others too...
- LLM APIs
- image generation
- text-to-speech
- transcription
- video generation
- MCP tools
- A2A agents
- AG-UI event streams
- vector databases
- search
- rerank
- moderation.
One package, one port, zero dependencies. Free.
https://t.co/qkZY51zJMq

[CopilotKit🪁 @CopilotKit]

Totally agree unit tests are critical, but curious if you've heard of AIMock? Mock infrastructure for AI application testing, and there's nothing like it that even comes close:
- LLM APIs
- image generation
- text-to-speech
- transcription
- video generation
- MCP tools
- A2A agents
- AG-UI event streams
- vector databases
- search
- rerank
- moderation.
One package, one port, zero dependencies. Open source. Free.
https://t.co/qkZY51zJMq