CopilotKit raises $27M to help developers deploy app-native AI agents

Most companies today add AI to their apps as a simple chatbot. Users type or dictate commands, and the AI bot tries to execute them. But this approach feels clunky. A text-based interface doesn’t always create smooth experiences, especially when you want to book an entire travel itinerary and have to sift through pages of text responses.

CopilotKit’s co-founders, Atai Barkai and Uli Barkai, believe this approach wastes the potential of AI agents and large language models. They argue the future lies in agents that live inside applications, understand user behavior, take actions, and display useful interfaces instead of returning long text blocks.

The Seattle-based startup has raised $27 million in a Series A round led by Glilot Capital, NFX, and SignalFire to bring this vision to market. The funding will support their enterprise toolkit built on top of their popular AG-UI protocol.

This funding represents a significant moment in the evolution of AI user interfaces. While most companies have focused on chat-based AI interactions, CopilotKit is betting on a more integrated approach where AI agents become native parts of applications. This shift could reshape how users interact with software across industries, from travel booking to business analytics.

CopilotKit’s AG-UI protocol is an open source standard that defines how AI agents connect to and communicate with user interfaces like web browsers or mobile apps. The protocol provides several key features:

• Streaming chat capabilities
• Front-end tool calls
• State sharing for human-in-the-loop functionality
• Dynamic UI generation based on context

The flexible user interface design is what sets CopilotKit apart. Instead of returning blocks of text, AI agents can generate interactive UIs using company-specific design components. For example, when a user asks for revenue breakdown by category, they get an interactive pie chart instead of a dense paragraph of numbers.

“The agent can reply to you, not just with blocks of text, but with interactive UIs that are defined by your own company,” CEO Atai Barkai explained. “All of your agents can, very trivially, speak to a UI and use these catalog of components and show that to users.”

Developers maintain full control over how much their AI agent can modify the interface. They can choose pixel-perfect precision or provide broad building blocks that AI assembles as needed.

The company is experiencing strong adoption across the tech ecosystem. AG-UI works alongside other widely used protocols like Model Context Protocol (MCP) and Agent2Agent (A2A). Major AI infrastructure providers including Google, Microsoft, Amazon, and Oracle now support the protocol, along with popular frameworks like LangChain, Mastra, PydanticAI, and Agno.

CopilotKit reports millions of weekly installs for its protocol and tools. A large portion of Fortune 500 companies are using the technology in production environments. Enterprise customers include Deutsche Telekom, Docusign, Cisco, and S&P Global.

To capitalize on this growing interest, CopilotKit is launching CopilotKit Enterprise Intelligence, a self-hostable offering that bundles infrastructure features for deploying agents within applications. This addresses two key enterprise requirements that Atai hears in almost every conversation: optionality and self-hosting capabilities.

The competitive landscape for enterprise AI tools is intense. Vercel’s open source AI SDK offers similar capabilities for building AI applications. Assistant-ui provides components for AI chat interfaces. OpenAI’s Apps SDK enables richer interfaces, though only within ChatGPT.

CopilotKit differentiates itself through a horizontal, enterprise-friendly approach rather than vertical integration. Instead of offering a full-stack AI platform, the company supports whatever agent framework, cloud provider, or backend enterprises already use.

“If there are two things we hear in almost every single enterprise conversation, enterprises want optionality and they want self-hosting,” Atai said. “Maybe they’re already using the Google, Amazon, Oracle, Microsoft, LangChain, Mastra stacks. They want optionality, and they want self-hosting, and these are two things that they don’t really get in the Vercel stack.”

The company faces a common tension for businesses built on open source infrastructure. They need to maintain AG-UI as a neutral standard while building a commercial product on top of it. The Barkais say AG-UI remains fully open, while CopilotKit’s commercial offerings add enterprise hardening to the open source foundation.

“Our strategy is to be the default choice in the ecosystem, and then to monetize the top enterprises,” said Uli Barkai, the startup’s head of growth. “So it’s very much in our interest that the open source is the best out there, and the 95% of users can just go build and get started without paying anyone or talking to anyone.”

CopilotKit currently employs about 25 people and plans to use the new funding to expand its team. The company’s approach could influence how the industry thinks about AI integration, moving from bolt-on chatbots to native agent experiences that feel more natural and powerful for users.