Products & Features
Build agent-native applications, connect to any agentic backend, and improve with Enterprise Intelligence.
Build. Connect. Improve.
Build agent-native applications with our best-in-class open source framework.
Connect to any agentic backend (framework, LLM, or mesh), and any protocol (AG-UI, MCP, or A2A).
Improve with our Enterprise Intelligence Platform for self-learning agents.
Run self-hosted or fully managed in CopilotKit Cloud.
CopilotKit Frontend SDKs
Open Source + Premium Features
Everything you need to build agent-native applications. Core features are open source under MIT license, with premium features available for teams that need full control.
Generative UI
AI agents dynamically generate and render UI components using MCP-Apps and A2UI - creating rich, interactive interfaces that go far beyond text chat.
Create engaging, visual experiences that adapt to each user's journey.
In-app Actions
Define actions that AI agents can execute directly within your application - from simple API calls to complex multi-step workflows.
Agents that actually do things, not just chat. Turn conversations into actions.
Realtime Context Awareness
Agents automatically stay aware of application state, user context, and environmental changes - no manual context management required.
Agents that understand context without explicit prompting.
Agent-App State Sync
Bidirectional state synchronization between agents and your application - agents read app state, app reacts to agent actions, everything stays in sync.
Seamless integration between conversational AI and your application.
React Client
Official CopilotKit client library for React applications with hooks, components, and full TypeScript support.
First-class React support. Build copilots the React way.
Fully Headless UI
Build completely custom copilot interfaces without any of CopilotKit's default UI components - full design freedom from the ground up.
Your copilot, your design. No compromises on user experience.
Observability Hooks
Programmatic access to all agent events, state changes, and performance metrics - build custom monitoring, logging, and analytics tailored to your needs.
Full visibility into agent behavior. Integrate with your existing observability stack.
Angular Client
Official CopilotKit client library for Angular applications with full TypeScript support and Angular-native patterns.
First-class Angular support. Build copilots the Angular way.
CopilotKit Inspector
Debug and inspect all agent-user interactions in real-time and historically. See exactly what your agents are doing, why they made decisions, and how they're performing.
X-ray vision into your agents. Debug with confidence.
Agentic Backend Connections
Any LLM, Any Framework, Any Protocol
Connect to any agentic backend - LLMs, agent frameworks, and protocols. AG-UI, MCP, A2A, LangChain, Google ADK, and more.
AG-UI Protocol
The Agent-User Interaction protocol enables seamless communication between AI agents and user interfaces - a standard for agentic frontend development.
Open standard for agent-UI communication. Build on a foundation that won't lock you in.
Agent Framework Integrations
Connect to any agent framework - LangChain, Google ADK, Mastra, and more. Use the best tools for your backend while CopilotKit handles the frontend.
Your backend, your choice. CopilotKit works with whatever you're using.
Protocol Support
Built-in support for MCP (Model Context Protocol) and A2A (Agent-to-Agent) communication - connect to the broader agent ecosystem.
Future-proof architecture built on open standards. Interoperability without lock-in.
Enterprise Intelligence Platform
Self-hosted or CopilotKit Cloud
Thread persistence, analytics, and continuous learning from users. Self-improving agents that get better over time.
Self-Hosted or Fully Managed
Deploy the Enterprise Intelligence Platform in your own infrastructure or let us manage it for you in CopilotKit Cloud. Same features, your choice of deployment.
Your infrastructure, your rules. Or let us handle it. The choice is yours.
Dashboard & IaC CI/CD
Visual dashboard for monitoring agent performance plus Infrastructure as Code for automated deployment and testing in CI/CD pipelines.
Production-grade operations. Deploy and monitor with confidence.
Thread History & Branching
End users can resume old conversations and branch from any point in conversation history - creating explorable, non-linear conversation experiences.
Conversations that persist and evolve over time. No more starting from scratch.
A/B Testing for Agentic Applications
Run experiments on agent behavior, prompts, flows, and UI with statistical analysis - data-driven optimization for AI applications.
Know what works. Optimize with data, not hunches.
Data Lakehouse & Analytics
Query and analyze all agent-user interactions across your application - a complete record of every conversation, action, and outcome.
Turn interactions into insights. Every conversation is learning data.
Self-improving Agents with RL
Agents that learn and improve from interactions without manual intervention - using reinforcement learning to optimize behavior over time.
Agents that learn from experience. Continuous improvement without manual work.
Enhance your applications
with powerful AI capabilities
Start with a single command
Copied