CopilotKit has raised $27 million in a Series A to help developers embed AI agents directly inside their applications. Instead of adding a chatbot sidebar, CopilotKit's open-source framework lets agents understand what users are doing, take actions within the app, and generate interactive UI components. The round was led by Glilot Capital, NFX, and SignalFire.
The Problem With AI Chatbots
Most companies today deploy AI as a text-based chatbot inside their apps. You type what you want. The bot generates a wall of text. You scan through it to find what you need. The experience feels clunky — especially for complex tasks like booking travel itineraries, analyzing data, or navigating enterprise workflows.
CopilotKit's founders argue that approach wastes what AI agents can actually do. Instead of returning text, an agent embedded through CopilotKit can generate interactive components — pie charts, forms, tables, calendars — that match the app's own design system. Ask for a revenue breakdown and you get a chart, not a paragraph.
The key difference is context. A chatbot sidebar does not know what you are looking at inside the app. CopilotKit's framework gives the agent awareness of the user's current state — what screen they are on, what data they are viewing, what actions are available. That context allows the agent to take relevant actions rather than generating generic responses.
AG-UI: The Open Protocol
CopilotKit's biggest contribution to the AI ecosystem is AG-UI — an open-source protocol that standardizes how AI agents connect to user interfaces. The protocol handles streaming chat, front-end tool calls, and state sharing between the agent and the application.
AG-UI works alongside two other widely adopted protocols: Anthropic's Model Context Protocol (MCP), which connects AI models to data sources, and Google's Agent2Agent (A2A), which enables communication between different AI agents. Together, these three protocols form an emerging infrastructure stack for agentic AI.
The adoption numbers are impressive. AG-UI is supported by Google, Microsoft, Amazon, and Oracle. Popular frameworks including LangChain, PydanticAI, and Mastra have integrated it. CopilotKit says it sees millions of installs per week and that a large portion of Fortune 500 companies use the protocol in production.
Enterprise Customers
CopilotKit counts Deutsche Telekom, Docusign, Cisco, and S&P Global as enterprise customers. The company also launched CopilotKit Enterprise Intelligence this week — a self-hostable offering that bundles infrastructure features for deploying agents within apps.
Self-hosting matters enormously for enterprise buyers. Companies in regulated industries — healthcare, finance, government — cannot send sensitive data through third-party cloud services. They need the ability to run AI tools on their own infrastructure. CopilotKit's self-hosted option addresses that requirement directly.
CEO Atai Barkai said the two things enterprises demand most are optionality and self-hosting. They want to choose their own agent framework, cloud provider, and backend. CopilotKit's horizontal approach — supporting whatever stack an enterprise already uses — differentiates it from vertically integrated alternatives like Vercel's AI SDK.
The Competition
The market for enterprise agent tooling is heating up. Vercel's open-source AI SDK offers similar capabilities for building AI applications. OpenAI's Apps SDK enables richer interfaces inside ChatGPT. And Microsoft Copilot recently made Agent Mode the default across Office — putting agentic capabilities directly into the world's most widely used enterprise software.
CopilotKit's advantage is neutrality. It does not lock developers into a specific AI provider, cloud platform, or framework. In an industry where enterprise buyers are increasingly wary of vendor lock-in, that flexibility is a meaningful selling point.
The company has about 25 employees and plans to use the new funding to grow its team.
The Bigger Picture
CopilotKit's $27 million raise reflects a broader shift in how the AI industry thinks about user experience. The first generation of AI products was built around chatbots. The next generation is embedding agents directly inside applications — giving them the ability to see, understand, and act within the context where users actually work.
The companies that build the infrastructure for that transition — the protocols, toolkits, and frameworks that connect agents to apps — will become essential plumbing for the agentic AI era. CopilotKit is betting that its open-source approach and enterprise-friendly positioning will make it the default choice.
