Playground
Interactive demos for every kit. Test features before implementing.
Core / Most Used
The essential building blocks for AI agents.
Structured Output
Type-safe structured data with Zod schemas.
Streaming
Real-time streaming with progress tracking.
Tool Calling
Define and execute AI tools with middleware.
Conversation State
Manage history with auto-summarization.
Step Reasoning
Chain-of-thought and tree-of-thought reasoning.
Multi-Step Planning
Break goals into executable step DAGs.
Observability & Debugging
Trace, log, and replay your agent's decisions.
UI / UX Agent Kits
React components purpose-built for AI agents.
Agent UI
Chat, streaming, and tool viewers.
Human-in-the-loop
Approval gates for critical actions.
Editable Responses
Fine-tune AI output before continuing.
Multimodal
Voice, images, and file handling.
State Visualization
Transparency into agent thought process.
Real-time & Collaboration
Multi-user agents and shared sessions.
Cognitive / Reasoning
Extend your agent's reasoning capabilities.
Tooling & Integration
Extend agent capabilities with external tools.
Execution & Control
Control workflows, retries, and budgets.
Safety & Governance
Production-ready security guardrails.
Model Abstraction
Route, fallback, and optimize model usage.
Evaluation & Testing
Benchmark, test, and detect regressions.
More Kits Coming
We have 48+ kits planned across 11 categories including Cognitive Reasoning, Safety, Observability, and more.