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

Many companies today offer AI merely as a chatbot embedded in their apps. Users type or dictate instructions, and the AI bot attempts to execute them. However, the experience often feels clunky. A text-based interface doesn't always deliver a smooth experience—for instance, when using a travel app to book a full itinerary, you might have to scroll through large amounts of text.

CopilotKit's founders believe this approach doesn't fully leverage the capabilities of AI agents and LLMs. Co-founders Atai Barkai (right in the image above) and Uli Barkai (left) see the future as enabling agents to operate within applications, understand user actions, take relevant steps, and present useful interfaces rather than merely returning lengthy text blocks.

CopilotKit's widely used AG-UI protocol addresses the first part of that vision. This open-source standard defines how AI agents connect and interact with user interfaces—such as web browsers or apps—offering features like streaming chat, front-end tool calls, and state sharing to support human-in-the-loop workflows. In essence, AG-UI provides developers with the framework and tools to embed AI agents inside their applications.

CopilotKit is also developing an enterprise toolkit built on AG-UI, adding support, self-hosted deployment options, and other essential features for companies considering integrating agents into their products. To bring this toolkit to market, the Seattle-based startup has raised $27 million in a Series A round led by Glilot Capital, NFX, and SignalFire, TechCrunch has exclusively learned.

The flexibility of the user interface is a key selling point. CEO Atai Barkai told TechCrunch that developers can use the startup's framework to define specifications and building blocks for dynamic interfaces, which an AI agent can then use to generate context-appropriate UIs.

“The agent can respond not just with text blocks but with interactive UIs designed by your own company,” Atai explained. “For instance, if a user asks for a revenue breakdown by category, instead of a dense, hard-to-read paragraph, they get a pie chart—your own pie chart design—that they can interact with [...] So all your agents can easily communicate with a UI, use the component catalog, and present that to users.”

Atai also noted that CopilotKit's toolkit gives developers complete control over how much their AI agent can modify the UI. They can choose to make the interface “pixel-perfect” or simply provide broad building blocks that the AI can assemble as needed.

The funding comes after a period of strong adoption for both AG-UI and CopilotKit. The protocol, which works alongside the widely used Model Context Protocol (MCP) and Agent2Agent (A2A) protocol, is now supported by major AI infrastructure providers such as Google, Microsoft, Amazon, and Oracle, as well as popular frameworks like LangChain, Mastra, PydanticAI, and Agno.

Atai said that CopilotKit and AG-UI—the company's strongest claim to ecosystem relevance—see millions of installs each week, and a large portion of Fortune 500 companies use the protocol and the startup's tools in production. Additionally, CopilotKit counts enterprise giants like Deutsche Telekom, Docusign, Cisco, and S&P Global as its customers.

To capitalize on this growing interest, the company is also launching CopilotKit Enterprise Intelligence, a self-hostable solution that bundles several infrastructure features for fully deploying agents within applications.

CopilotKit faces intense competition in the enterprise agent tools market. Cloud platform Vercel's open-source AI SDK helps developers build AI applications with similar capabilities, while Assistant-ui provides components for creating AI chat interfaces. Meanwhile, OpenAI's Apps SDK also offers a way to build richer interfaces, but only within ChatGPT.

Atai argues that CopilotKit differs from these offerings because it takes a horizontal, enterprise-friendly approach rather than a vertically integrated one. Instead of providing a full-stack AI platform, CopilotKit aims to support whatever agent framework, cloud provider, or backend an enterprise already has.

“If there are two things we hear in almost every enterprise conversation, it's that they want optionality and self-hosting,” he said. “They might already be using the Google, Amazon, Oracle, Microsoft, LangChain, or Mastra stacks. They want optionality and self-hosting—two things they don’t really get from the Vercel stack.”

Maintaining that open positioning will be crucial. Companies that build on their own open-source infrastructure often face a tension: they want their technology to remain a neutral standard while also building a business on top of it. However, Atai said that AG-UI is a fully open protocol, and CopilotKit's commercial product is designed to strengthen the open-source stack for enterprises, not replace it.

“They are very complementary. Our strategy is to become the default choice in the ecosystem and then monetize the top enterprises,” added Uli, the startup's head of growth. “So it's in our best interest that the open source is the best available, and that 95% of users can just build and get started without paying or talking to anyone.”

The company currently has about 25 employees and plans to use the new funding to expand its team.