PricingPricing

Build your first copilot in minutes

Choose the CopilotKit Premium plan that works for you based on your stage of development and number of users

Developer

Individual

Free

  • Up to 100 Monthly Active Users
  • Core CopilotKit
  • Agent Framework Integrations
  • Premium Features
  • Early Access to New Features
  • CopilotKit Cloud SaaS
  • Discord community support

Pro

Teams

$249

/ month
  • Up to 1000 Monthly Active Users

    Everything in Developer plus:

  • Team accounts
  • Dedicated Slack support

Enterprise

Scale

Contact Us

  • Custom plans for more than 1000 Monthly Active Users

    Everything in Pro plus:

  • SSO
  • On-prem Cloud
  • Internationalization
  • Accessibility support
  • Onboarding consultation, dedicated Slack support and SLA

CopilotKit Open Source

Access core CopilotKit features on an always free, self-hosted framework.

Trusted by developers at companies worldwide:
Cisco Logo
Deloitte Logo
Trip Advisor Logo
Tavily Logo
Lang Chain Logo
Mastra Logo
Function Logo
Cisco Logo
Deloitte Logo
Trip Advisor Logo
Tavily Logo
Lang Chain Logo
Mastra Logo
Function Logo
Integration
Seamless integration with Multi-Agent Frameworks

CopilotKit's feature rich SDK includes custom-built adapters for the top AI Agent frameworks

illustration
Available for Preview - Contact Us

CopilotKit's Agent-User Platform

Everything you need to build, deploy, and scale user-facing agents and agentic applications—with built-in analytics, security, and Reinforcement Learning

The Essentials

Core capabilities you need to build user-facing agentic applications

Use any LLM

Seamlessly integrate with any large language model - OpenAI, Anthropic, local models, and more

Drop-in agents, or AG-UI

Pre-built agent templates that just work, or use any AG-UI compatible agent framework

Threads

Persistent conversation threads that maintain context across interactions

Voice support

Native voice input and output capabilities for natural conversational experiences

Hosting and CI/CD

Production-ready hosting with automated deployment pipelines and monitoring

Advanced debugging tools

Comprehensive debugging suite to troubleshoot and optimize agentic apps

The AG-UI Cockpit

Request Access

Understand the interaction between your agentic applications and your users

The AG-UI Cockpit is the dashboard for product analytics that fit agentic apps. See what users are trying to do with your copilots, the real paths they take across conversations, generated UI, and page visits, and where they succeed or get stuck. Turn messy interaction data into clear, actionable insights.

See the full journey

Trace each user thread end to end, across chat turns, generated components, form fills, and page hops.
Tie entry points to outcomes: submission, export, handoff, or abandon.

Aggregate metrics that matter

Success rate and effectiveness, time to value, turns to completion, correction rate, sentiment.
Rollups by intent, cohort, entry point, and page context.

Agentic A/B testing

Run true agent-level experiments: compare agent variants or UI options.
Measure impact on end-to-end user flows and business outcomes, then roll out winners with confidence.

Private by design

Boundary-only capture with redaction, self-host option, works with your stack.

The AG-UI Shield

Request Access

Protect your users, organization, and brand from unsafe or malicious behavior

The AG-UI Shield delivers security at the agent-user boundary. Stop prompt injections, prevent data leaks, and enforce policy on both inputs and outputs. One checkpoint for everything, so nothing slips through.

Why it matters

Guardrails scattered across agents and sub-agents are error prone. The AG-UI Shield applies controls at the single place every message, file, UI action, tool response, and state update must pass.

What it covers

Chats and multi-turn threads, shared state between agent and app, generated UI, files and uploads, tool calls and responses.

What it enforces

Injection control: Block attempts to steer the agent into actions it should not take.
Data loss prevention: Stop confidential data from tools, connectors, or memory before it reaches users.
Brand and policy: Keep tone and claims on brand, within policy, every time.

How it operates

Declarative policies, real-time actions (block, redact, warn, escalate), full audit logs. Model and framework agnostic, private by design, self-host option.

The AG-UI Reinforcement Learning (RL) Layer

Request Access

Build self-improving agents from real user interactions. The AG-UI RL Layer converts agent-user interactions into reinforcement learning signals at the boundary, so your copilots get smarter, faster, and safer with every use.

Core capabilities

Agent-user interaction capture at the boundary
Record edits, approvals, human-in-the-loop steps, retries, tool calls, outputs, and shared-state updates as structured, replayable events.
Learning Containers
Scope learning to a user, a team, an organization, or your full user base.
Privacy and control
Self-hostable, boundary-only capture, model and framework agnostic. Your data stays in your environment.

Adaptation tracks

In-context RL
Apply past reinforcement signals at generation time.
RL Dataset Export
Produce RL datasets you can fine-tune on your own stack.
Closed-loop fine-tuning
Connect to leading model providers or your infra to create reinforced model layers.

Outcomes

Higher task success and quality
Real interaction feedback lifts accuracy on the workflows that matter.
Shorter development time
Agents learn from usage, so you spend less time hand-tuning subtle paths at build time.
Are you ready?

Enhance your applications
with powerful AI capabilities