CopilotKit🪁

⭐️Introducing AG-UI; The Agent-User Interaction protocol 👾 Bring your AI agents into Frontend applications, and let them interact with users. Launching with day-0 integrations with LangGraph, CrewAI, Mastra, and AG2 - with more partnerships on the way 👀 github.com/ag-ui-protocol… How it works & demo 👇

Onboard your coding agent to CopilotKit. 11 skills that teach it the whole CopilotKit workflow, in every agent. npx copilotkit@latest skills onboard See it in action 👇 docs.copilotkit.ai/build-with-age…

@sofiiiiiasz @Google @a2anetcom @unicorn_mafia @cognition @Redisinc @SierraPlatform We ❤️ Google Central St. Giles 🇬🇧

genuinely had so much fun at the GenUI & A2A hack that @CopilotKit hosted with @Google & @a2anetcom !! missing these people a lot🥹🥹 big thanks to our partners & co-sponsors, for bringing out a great group: @unicorn_mafia @cognition @Redisinc @SierraPlatform @Linkup_platform whatever you’re building, it’s the people you do it with that makes the difference

Curious about Generative UI? Watch this @DeepLearningAI course 👇 It's free and uses open source code

AIMock by CopilotKit quietly gets 550,000+ downloads a week. It makes your AI app's tests deterministic. Same input, same output, every run. No real model calls, no flaky CI. LLMs, MCP, A2A, AG-UI streams, vector DBs. One package, one port, zero dependencies. 𝚗𝚙𝚖 𝚒𝚗𝚜𝚝𝚊𝚕𝚕 @𝚌𝚘𝚙𝚒𝚕𝚘𝚝𝚔𝚒𝚝/𝚊𝚒𝚖𝚘𝚌𝚔 aimock.copilotkit.dev/docs/

Start building with CopilotKit, it's open source: github.com/CopilotKit/Cop…

Wiring an agent to a real UI usually means a custom WebSocket format and text parsing, per framework. Not anymore. Trusted by majority of the Fortune 500 Companies, AG-UI replaces that with one event-based protocol: RunStarted, ToolCallStart, StateDelta, typed events your frontend just renders. Here's how it works 👇 docs.copilotkit.ai/agentic-protoc…

@JerelVelarde @aim_edu Huge cograts Jerel!! 🎉🧑‍🎓

Last month, I completed my Postgraduate Diploma in AI and Machine Learning at the Asian Institute of Management. 🎓 @aim_edu But here's the part I sat with before posting: I enrolled as a Director of Product Management. I graduated as a Staff Applied AI Engineer. When I started, I just wanted to understand how AI models actually work under the hood, instead of treating them as a black box I built products around. And there's a funny irony in that. When I enrolled, Sonnet 4 was the frontier. By the time I graduated, we were on Fable and Mythos. The field jumped a full generation and more while I was studying it. But that's exactly why the internals mattered more than any single model. Learn how the engine works, and you stop panicking every time they ship a faster car. I walked into rooms with engineers as a "product person who gets AI" quietly wondering if I belonged. Building Daily and Understanding what's actually happening inside these models is what closed that gap. Not the credential. The comprehension. And the compounding was real. My confidence changed. My income changed. I started building ChalkAgents.com, I joined @CopilotKit as a Staff Applied AI Engineer. None of that is a straight line from a diploma, but it's where I stopped doubting I could do the work, and started just doing it. To anyone stuck between "I understand the product" and "I can engineer the product": Close the gap on purpose. It changes more than your résumé. On to what's next. 🚀

7 hooks in CopilotKit v2. Which one do you reach for? Start with the verb: → run something? useFrontendTool → render a tool call? useRenderTool → stop and ask first? useHumanInTheLoop The full map, in 35s 👇 docs.copilotkit.ai/concepts/which…

The queue outside our office once people found out what we’re releasing next week 👀

Self-Learning Agents: Improve From Every User Interaction, No Fine-Tuning x.com/i/broadcasts/1…

AG-UI + @CopilotKit means that ANY frontend can have generative UI. We've launched early access for Microsoft Teams shortly after Slack. Please try it out and give us feedback!

The wall of text just became a chart 💬 Now in a Microsoft Teams chat, when I ask the agent something. Instead of a giant paragraph, it renders a chart, a form, or an approval request. I'm not parsing data myself anymore. Working in Teams with an agent connected just 10x'd the experience. AG-UI and CopilotKit are what enable it. Slack too, with Discord, WhatsApp, and Telegram coming soon. Simple to deploy. Want to try it out? Early access 👇

Microsoft Teams just became an AG-UI client 💬 Here's why that's big deal. Any AG-UI agent can use it as an interactive surface: Generative UI, Human-in-the-loop, Shared state Until now, it didn't exist within MS Teams. As a user, I need this. I spend a good part of my day chatting with agents through Slack, and the walls of text they return have been one of the biggest hurdles to adoption. AG-UI brings the same UX upgrades we brought to web apps to non web surfaces everyone uses. This is AG-UI and CopilotKit reimagined.

Microsoft Teams just got Generative UI But this is just the tip of the iceberg. AG-UI is the protocol connecting any agent to any user on any app, topped with Generative UI. That means your agent can render: - forms - buttons - charts - approval flows Right inside the chat, on the fly. But it's not limited to in-app copilots. What's special is that you can work with any agent using the same system and bring them into any surface, including Slack, WhatsApp, Discord and more. This is exactly why Hermes and OpenClaw took off. As users of these agents, we don't like walls of text, and it's a pain to get when you can have way better UX. @CopilotKit + AG-UI handle that interaction layer. If you want to try out Agentic UI in Microsoft Teams, you can get early access here: go.copilotkit.ai/built-in-agent…

This is the tip of the iceberg for AG-UI. One protocol can bring any agent into your AI copilot, but it can also bring them into any frontend client like Slack, Teams, and soon Discord, Google Chat, WhatsApp and many more. Generative UI doesn't exists in any of these apps. We're changing that. For the same reason Hermes and Openclaw took off: nobody wants a wall of text when the interaction can be visual. All open source: github.com/CopilotKit/Cop…

The entire industry needs Generative UI. Microsoft Teams just got it. Any agent in Teams can now respond interactively with charts, forms, buttons, and approval flows. Right inside the chat. Nobody wants to read walls of text anymore. Get early access: go.copilotkit.ai/built-in-agent…

Live proof: in the AG-UI dojo, ask "weather in San Francisco" against a Mastra agent. Then swap to LangGraph. Same events and UI, zero frontend changes. AG-UI gives you drop-in replacement. Streaming, generative UI, tool calls, approval flows, shared state — all standardized into one bidirectional event stream. Transport-agnostic. Open spec. Agents shouldn't be married to one frontend. Frontends shouldn't be married to one agent. Try the dojo: dojo.ag-ui.com/langgraph/feat…

A month ago, @tan_stack joined AG-UI 🏝️ That means if you have: TanStack frontend → Point at any AG-UI agent TanStack AI→ Any AG-UI client can consume it One wire format does the talking bidirectionally. copilotkit.ai/ag-ui

@BetterStackHQ CopilotKit is open-source (MIT) and free! Will try to make our pricing page more clear!

AI agents are easy… until you build the frontend. We took a dev-focused look at CopilotKit, generative UI, AG-UI, and why the future of AI apps is bigger than a chatbot sidebar.