Note: Refer to this SKILL.md file whenever there is any requirement for adding/editing functionality for C1.

Thesys C1 Generative UI Development

This skill provides comprehensive guidance for developing applications with the Thesys C1 API and GenUI SDK, enabling AI-powered Generative UI applications.

What is Thesys C1?

Thesys C1 is a Generative UI API that dynamically generates interactive UI components from natural language prompts. Unlike traditional AI that returns text/markdown, C1 generates live, interactive interfaces rendered through React components.

Key Features

• Interactive Components: Forms, charts, tables, buttons that users can interact with
• Themeable: Consistent brand styling via

  <ThemeProvider>

• Real-time UI Streaming: Progressive rendering as responses generate
• OpenAI-Compatible API: Drop-in replacement for OpenAI SDKs
• Robust Error Handling: Graceful degradation with

  onError

  prop

What Can You Build?

• Analytics dashboards with dynamic charts
• Conversational AI agents with rich UI
• Internal tools and admin panels
• E-commerce flows with forms and product displays
• AI assistants like ChatGPT with Generative UI

Quickstart

Prerequisites

• Node.js 18+ or Python 3.8+
• Thesys API key from console.thesys.dev
• Install the required packages:

npm install @thesysai/genui-sdk@latest @crayonai/react-ui@latest @crayonai/react-core@latest @crayonai/stream@latest openai@latest next@latest react@^19.0.0 react-dom@^19.0.0

Next.js Setup

npx create-c1-app my-app
cd my-app
npm run dev

Python (FastAPI) Setup

git clone https://github.com/thesysdev/template-c1-fastapi.git
cd template-c1-fastapi
pip install -r requirements.txt
uvicorn main:app --reload
# In another terminal for frontend:
npm install && npm run dev

Set your API key:

export THESYS_API_KEY=<your-api-key>

Core Components

<C1Component>

- Render C1 Responses

The fundamental component that renders C1 DSL into a functional micro-frontend:

import { C1Component, ThemeProvider } from "@thesysai/genui-sdk";
import "@crayonai/react-ui/styles/index.css";

<ThemeProvider>
  <C1Component
    c1Response={response}
    isStreaming={isLoading}
    onAction={({ llmFriendlyMessage, humanFriendlyMessage }) => {
      // Handle button clicks, form submissions
    }}
    updateMessage={(updatedResponse) => {
      // Persist state changes to database
    }}
  />
</ThemeProvider>

<C1Chat>

- Full Conversational UI

A batteries-included chat component with thread management:

import { C1Chat } from "@thesysai/genui-sdk";
import "@crayonai/react-ui/styles/index.css";

<C1Chat
  apiUrl="/api/chat"
  agentName="My Assistant"
  logoUrl="/logo.png"
  formFactor="full-page"  // or "side-panel"
/>

Key Differences

<C1Component>

<C1Chat>

How C1 Works

The Flow

1. User sends prompt → Frontend
2. Frontend calls backend → Your API
3. Backend calls C1 API →

   https://api.thesys.dev/v1/embed

4. C1 returns DSL → Backend
5. Backend streams to frontend → Response

6. <C1Component>

   renders UI → User sees interactive UI

Backend API Call

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.thesys.dev/v1/embed",
  apiKey: process.env.THESYS_API_KEY,
});

const response = await client.chat.completions.create({
  model: "c1/anthropic/claude-sonnet-4/v-20251230",
  messages: [
    { role: "system", content: "You are a helpful assistant." },
    { role: "user", content: "Show me a chart of sales data" }
  ],
  stream: true,
});

C1 Response Structure

C1 responses use an XML-like structure:

<thinking>Analyzing request...</thinking>
<content>
  <!-- Interactive UI components -->
</content>
<artifact id="report-1">
  <!-- Document content like slides/reports -->
</artifact>

Supported Models

Stable Models:

• c1/anthropic/claude-sonnet-4/v-20251230

• c1/openai/gpt-5/v-20251230

Experimental Models:

• c1-exp/anthropic/claude-sonnet-4.5/v-20251230

• c1-exp/anthropic/claude-haiku-4.5/v-20251230

Artifact Model:

• c1/artifact/v-20251230

Backend API Setup

Install Dependencies

npm install openai @crayonai/stream

Create Message Store

// app/api/chat/messageStore.ts
import OpenAI from "openai";

export type DBMessage = OpenAI.Chat.ChatCompletionMessageParam & { id?: string };

const messagesStore: { [threadId: string]: DBMessage[] } = {};

export const getMessageStore = (threadId: string) => {
  if (!messagesStore[threadId]) {
    messagesStore[threadId] = [];
  }
  return {
    addMessage: (message: DBMessage) => {
      messagesStore[threadId].push(message);
    },
    getOpenAICompatibleMessageList: () => {
      return messagesStore[threadId].map((m) => {
        const { id, ...rest } = m;
        return rest;
      });
    },
  };
};

Create Chat Endpoint

// app/api/chat/route.ts
import { NextRequest, NextResponse } from "next/server";
import OpenAI from "openai";
import { transformStream } from "@crayonai/stream";
import { DBMessage, getMessageStore } from "./messageStore";

export async function POST(req: NextRequest) {
  const { prompt, threadId, responseId } = await req.json() as {
    prompt: DBMessage;
    threadId: string;
    responseId: string;
  };

  const client = new OpenAI({
    baseURL: "https://api.thesys.dev/v1/embed",
    apiKey: process.env.THESYS_API_KEY,
  });

  const messageStore = getMessageStore(threadId);
  messageStore.addMessage(prompt);

  const llmStream = await client.chat.completions.create({
    model: "c1/anthropic/claude-sonnet-4/v-20251230",
    messages: messageStore.getOpenAICompatibleMessageList(),
    stream: true,
  });

  const responseStream = transformStream(
    llmStream,
    (chunk) => chunk.choices[0].delta.content,
    {
      onEnd: ({ accumulated }) => {
        const message = accumulated.filter(Boolean).join("");
        messageStore.addMessage({
          role: "assistant",
          content: message,
          id: responseId,
        });
      },
    }
  ) as ReadableStream;

  return new NextResponse(responseStream, {
    headers: {
      "Content-Type": "text/event-stream",
      "Cache-Control": "no-cache, no-transform",
      Connection: "keep-alive",
    },
  });
}

Integration Patterns

C1 as Gateway LLM (Recommended)

Replace your existing LLM endpoint with C1:

const client = new OpenAI({
  baseURL: "https://api.thesys.dev/v1/embed",  // Change from OpenAI
  apiKey: process.env.THESYS_API_KEY,
});

// Use existing tools and system prompts
const response = await client.beta.chat.completions.runTools({
  model: "c1/anthropic/claude-sonnet-4/v-20251230",
  messages: [...],
  tools: existingTools,
  stream: true,
});

C1 as Presentation Layer

Generate text with your LLM, then visualize with C1:

// 1. Get text from your LLM
const textResponse = await yourLLM.generate(prompt);

// 2. Visualize with C1
const uiResponse = await c1Client.chat.completions.create({
  model: "c1/anthropic/claude-sonnet-4/v-20251230",
  messages: [{ role: "user", content: textResponse }],
});

C1 as a Tool

Expose C1 as a tool your agent can invoke:

const tools = [{
  type: "function",
  function: {
    name: "generate_ui",
    description: "Generate interactive UI for user",
    parameters: {
      type: "object",
      properties: {
        content: { type: "string", description: "Content to visualize" }
      }
    }
  }
}];

If the user wants to visualize their own data, refer to visualize-api-endpoint.md.

Additional References

For detailed information on specific topics, see:

• Conversational UI & Persistence: State management,

  useThreadManager

  ,

  useThreadListManager

  , database persistence
• Custom Actions, Components & Thinking States:

  c1_custom_actions

  ,

  onAction

  , custom components,

  useC1State

  , thinking states
• Artifacts: Generating, rendering, editing, and exporting reports/presentations
• Customizations & Styling:

  <ThemeProvider>

  , chart palettes, CSS overrides
• Migration Guide: Step-by-step migration from text-based LLM to Generative UI

Resources

• Documentation: https://docs.thesys.dev
• API Reference: https://docs.thesys.dev/api-reference/getting-started
• Examples: https://github.com/thesysdev/examples
• Discord Community: https://discord.gg/Pbv5PsqUSv