CEO @CopilotKit🪁 | ex-@Meta | Author of PsyPhy (experimentally testable physics theory of consciousness) | Physics BA & MSc @Penn
- Tweets 3,606
- Following 2,961
- Followers 10,039
- Likes 12,874
Pinned Tweet
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.
60
58
459
353,123
In 2027, every app will ship Intelligent UI.
We open sourced it👇
Open Intelligent UI, built with @CopilotKit
(#5 Trending, 4k Github stars)
CopilotKit makes Intelligent UI possible in ANY app, and works with:
- Any LLM
- Any agent framework
- Your own components and design system
- Every kind of generative UI
We built it around flexibility and no lock-in.
👉 Try Open Intelligent UI: github.com/CopilotKit/OpenIn…
CopilotKit is the product stack for agents with nearly 40k stars and 16M+ downloads. Repo at the end.
See Intelligent UI examples below
5
2
16
1,228
It can pick the format for you.
Text when that’s enough.
A timeline, chart or table when that helps more.
1
2
40
If you want to build full-stack agentic apps, check out the @CopilotKit repo:
github.com/CopilotKit/Copilo…
3
42
🎉 OpenDots now works with @composio
Your personal AI coworkers can now work across Gmail, GitHub, Notion, Linear and 1,000+ more apps.
Chat, call, or even @ them in Slack.
Contributed by the Composio team 👏
5
2
47
2,843
Use @CopilotKit's OpenMuse with E2B to give your agent a full desktop computer for last-minute meeting prep.
Have it do the legwork: write Python to pull five years of revenue and net income from SEC EDGAR, save the data as a CSV, and open the charts in a browser.
2
5
37
2,195
Clone OpenMuse and get started:
github.com/CopilotKit/openmu…
11
581
Atai Barkai retweeted
Claude Managed Agents can power many different UIs.
Our new cookbook pairs Claude Managed Agents with AG-UI, @CopilotKit's open Agent-User Interaction Protocol.
105
142
2,405
338,069
Upvote here: producthunt.com/products/cop…
57
Atai Barkai retweeted
Wow, we had HUNDREDS of people sign up for our Continual Learning webinar this morning and had to keep hundreds on the waitlist.
Super engaged audience & really awesome questions and feedback. Thrilled to see this product getting real world interest, we're really proud of it.
INTRODUCING: CopilotKit CLiF
A platform for Continual Learning from Full-stack Trajectories.
Teach your agents how work gets done by turning the work users & employees do into:
→ Exportable datasets (of trajectories)
→ Evals (bring into platform of choice)
→ Skills
→ Fine-tuned models
Your users already do the exact tasks you're trying to teach your agent, every day.
Yet months of evals and engineering leave out your most important dataset...
Product trajectories.
AG-UI already sees every step your agent takes. Now it can also securely capture what your users do: what was on their screen, the buttons they clicked, the server requests they triggered.
Together that's a full-stack trajectory.
Your people already do the work.
Now agents can learn it from them.
Compatible with your existing observability and evals stack. Fully self-hostable.
@CopilotKit CLiF is in private alpha.
Book a demo / Get early access: go.copilotkit.ai/get-clif
More below 👇
2
8
529
Atai Barkai retweeted
CopilotKit just released CLiF (Continual Learning from Full-Stack Trajectories), a platform that teaches AI agents from what users actually do in the product, not just from what the agent did.
The core idea is the full-stack trajectory: the agent trace (runs, tool calls, state changes) and the user's product actions (screen, clicks, server requests) captured as one ordered timeline over AG-UI.
CLiF analyzes those trajectories, surfaces Insights like repeated corrections and friction users quietly work around, and proposes skills as plain SKILL.md files. A human approves every skill. Published skills reach the agent without a redeploy, and can ship with an MCP server so external agents like Claude follow them too.
The same trajectories also become production-based evals, fine-tuning datasets for smaller open models, and raw exports for any observability stack. Fully self-hostable on Kubernetes or AWS ECS/Fargate......
Full analysis: marktechpost.com/2026/10/15/…
CLiF: copilotkit.ai/clif?utm_sourc…
Technical details: copilotkit.ai/blog/introduci…
@CopilotKit
7
10
29
55,403
CLiF 🤝 Google ADK
Here's a practical how-to guide for adding Automatic Learning to your ADK Agent.
(Optimized for coding assistants)
CLiF turns real user work into:
> Product insights
> Approvable skills
> A self-improving system
Get started: copilotkit.ai/blog/google-ad…
INTRODUCING: CopilotKit CLiF
A platform for Continual Learning from Full-stack Trajectories.
Teach your agents how work gets done by turning the work users & employees do into:
→ Exportable datasets (of trajectories)
→ Evals (bring into platform of choice)
→ Skills
→ Fine-tuned models
Your users already do the exact tasks you're trying to teach your agent, every day.
Yet months of evals and engineering leave out your most important dataset...
Product trajectories.
AG-UI already sees every step your agent takes. Now it can also securely capture what your users do: what was on their screen, the buttons they clicked, the server requests they triggered.
Together that's a full-stack trajectory.
Your people already do the work.
Now agents can learn it from them.
Compatible with your existing observability and evals stack. Fully self-hostable.
@CopilotKit CLiF is in private alpha.
Book a demo / Get early access: go.copilotkit.ai/get-clif
More below 👇
6
2
26
2,352
Atai Barkai retweeted
The MOST expensive problem today is evals.
Now you can use people's real work to improve your entire product.
The AG-UI protocol captures everything:
→ what the agents did
→ what the users did
AG-UI is the open protocol between agents and user-facing apps, adopted by Google, Microsoft, AWS and more, and it sees both sides of every task.
@CopilotKit (the company behind AG-UI) is using that interaction data to make agents learn from how the product is actually used.
It's called CLiF-
Continual Learning from Fullstack Trajectories.
Over time, CLiF finds patterns across sessions and proposes:
> datasets of fullstack trajectories
> evals
> skills
> and fine-tuned models
All of it is yours to export, and you can self-host it inside your own VPC.
12
10
65
3,603
New in AG-UI: Capture everything your users do
Their clicks, screens, requests...
Then use CLiF to continually improve your product by generating skills & evals.
👉 Get early access to this: go.copilotkit.ai/get-clif
INTRODUCING: CopilotKit CLiF
A platform for Continual Learning from Full-stack Trajectories.
Teach your agents how work gets done by turning the work users & employees do into:
→ Exportable datasets (of trajectories)
→ Evals (bring into platform of choice)
→ Skills
→ Fine-tuned models
Your users already do the exact tasks you're trying to teach your agent, every day.
Yet months of evals and engineering leave out your most important dataset...
Product trajectories.
AG-UI already sees every step your agent takes. Now it can also securely capture what your users do: what was on their screen, the buttons they clicked, the server requests they triggered.
Together that's a full-stack trajectory.
Your people already do the work.
Now agents can learn it from them.
Compatible with your existing observability and evals stack. Fully self-hostable.
@CopilotKit CLiF is in private alpha.
Book a demo / Get early access: go.copilotkit.ai/get-clif
More below 👇
2
1
69
5,871
Atai Barkai retweeted
Man am I stoked for this release, we're making your agents smarter as your users consume them even if they don't directly interact with the chat window.
We let your agents learn about painful user journeys and improve themselves to make them smoother 🔥
INTRODUCING: CopilotKit CLiF
A platform for Continual Learning from Full-stack Trajectories.
Teach your agents how work gets done by turning the work users & employees do into:
→ Exportable datasets (of trajectories)
→ Evals (bring into platform of choice)
→ Skills
→ Fine-tuned models
Your users already do the exact tasks you're trying to teach your agent, every day.
Yet months of evals and engineering leave out your most important dataset...
Product trajectories.
AG-UI already sees every step your agent takes. Now it can also securely capture what your users do: what was on their screen, the buttons they clicked, the server requests they triggered.
Together that's a full-stack trajectory.
Your people already do the work.
Now agents can learn it from them.
Compatible with your existing observability and evals stack. Fully self-hostable.
@CopilotKit CLiF is in private alpha.
Book a demo / Get early access: go.copilotkit.ai/get-clif
More below 👇
4
14
1,073