Error Handling

The Decart AI SDK provides structured error handling with specific error codes for different failure scenarios. This guide shows you how to handle errors gracefully.

Error Types

All SDK errors follow the DecartSDKError type:

type DecartSDKError = {
  code: string;
  message: string;
  data?: Record<string, unknown>;
  cause?: Error;
};

Error Codes

The SDK defines these error codes:

export const ERROR_CODES = {
  // General errors
  INVALID_API_KEY: "INVALID_API_KEY",
  INVALID_BASE_URL: "INVALID_BASE_URL",
  PROCESSING_ERROR: "PROCESSING_ERROR",
  INVALID_INPUT: "INVALID_INPUT",
  INVALID_OPTIONS: "INVALID_OPTIONS",
  MODEL_NOT_FOUND: "MODEL_NOT_FOUND",
  
  // Queue API errors
  QUEUE_SUBMIT_ERROR: "QUEUE_SUBMIT_ERROR",
  QUEUE_STATUS_ERROR: "QUEUE_STATUS_ERROR",
  QUEUE_RESULT_ERROR: "QUEUE_RESULT_ERROR",
  JOB_NOT_COMPLETED: "JOB_NOT_COMPLETED",
  
  // Token errors
  TOKEN_CREATE_ERROR: "TOKEN_CREATE_ERROR",
  
  // WebRTC errors (real-time)
  WEBRTC_WEBSOCKET_ERROR: "WEBRTC_WEBSOCKET_ERROR",
  WEBRTC_ICE_ERROR: "WEBRTC_ICE_ERROR",
  WEBRTC_TIMEOUT_ERROR: "WEBRTC_TIMEOUT_ERROR",
  WEBRTC_SERVER_ERROR: "WEBRTC_SERVER_ERROR",
  WEBRTC_SIGNALING_ERROR: "WEBRTC_SIGNALING_ERROR",
};

Basic Error Handling

Process API (Synchronous)

import { createDecartClient, models, ERROR_CODES } from "@decartai/sdk";

try {
  const client = createDecartClient({ apiKey: "your-api-key" });
  
  const blob = await client.process({
    model: models.image("lucy-pro-t2i"),
    prompt: "a beautiful sunset",
  });
  
  console.log("Image generated successfully");
} catch (error) {
  // Type guard to check if it's a DecartSDKError
  if (typeof error === 'object' && error !== null && 'code' in error) {
    const sdkError = error as { code: string; message: string };
    
    switch (sdkError.code) {
      case ERROR_CODES.INVALID_API_KEY:
        console.error("Invalid API key. Please check your credentials.");
        break;
      case ERROR_CODES.INVALID_INPUT:
        console.error("Invalid input:", sdkError.message);
        break;
      case ERROR_CODES.PROCESSING_ERROR:
        console.error("Processing failed:", sdkError.message);
        break;
      default:
        console.error("Unexpected error:", sdkError.message);
    }
  } else {
    console.error("Unknown error:", error);
  }
}

Queue API (Async Jobs)

try {
  const job = await client.queue.submit({
    model: models.video("lucy-pro-t2v"),
    prompt: "a flying bird",
  });
  
  console.log("Job submitted:", job.job_id);
  
  // Poll for status
  const status = await client.queue.status(job.job_id);
  console.log("Job status:", status.status);
  
  // Get result (only when completed)
  if (status.status === "completed") {
    const blob = await client.queue.result(job.job_id);
    console.log("Video ready");
  }
} catch (error) {
  if (typeof error === 'object' && error !== null && 'code' in error) {
    const sdkError = error as { code: string; message: string; data?: any };
    
    switch (sdkError.code) {
      case ERROR_CODES.QUEUE_SUBMIT_ERROR:
        console.error("Failed to submit job:", sdkError.message);
        if (sdkError.data?.status) {
          console.error("HTTP status:", sdkError.data.status);
        }
        break;
      case ERROR_CODES.JOB_NOT_COMPLETED:
        console.error("Job not ready yet:", sdkError.message);
        console.error("Current status:", sdkError.data?.currentStatus);
        break;
      default:
        console.error("Queue error:", sdkError.message);
    }
  }
}

Real-Time Error Handling

For real-time connections, handle errors via the event emitter:

import { createDecartClient, type DecartSDKError, ERROR_CODES } from "@decartai/sdk";

const client = createDecartClient({ apiKey: "your-api-key" });

try {
  const realtimeClient = await client.realtime.connect(stream, {
    model: models.realtime("mirage_v2"),
    onRemoteStream: (transformedStream) => {
      console.log("Receiving video stream");
    },
  });

  // Subscribe to connection errors
  realtimeClient.on("error", (error: DecartSDKError) => {
    console.error("WebRTC error:", error.code, error.message);
    
    switch (error.code) {
      case ERROR_CODES.WEBRTC_WEBSOCKET_ERROR:
        console.error("WebSocket connection failed");
        // Try reconnecting or show user message
        break;
        
      case ERROR_CODES.WEBRTC_ICE_ERROR:
        console.error("ICE connection failed (network issue)");
        // Check network/firewall settings
        break;
        
      case ERROR_CODES.WEBRTC_TIMEOUT_ERROR:
        console.error("Connection timed out");
        if (error.data?.timeoutMs) {
          console.error(`Timeout after ${error.data.timeoutMs}ms`);
        }
        // Retry connection
        break;
        
      case ERROR_CODES.WEBRTC_SERVER_ERROR:
        console.error("Server error:", error.message);
        // Server-side issue, wait and retry
        break;
        
      default:
        console.error("Unhandled WebRTC error");
    }
  });

  // Monitor connection state
  realtimeClient.on("connectionChange", (state) => {
    console.log("Connection state:", state);
    
    if (state === "disconnected") {
      console.log("Connection lost");
    } else if (state === "reconnecting") {
      console.log("Attempting to reconnect...");
    }
  });
  
} catch (error) {
  console.error("Failed to establish connection:", error);
}

Diagnostic Events

For real-time connections, you can also listen to diagnostic events:

realtimeClient.on("diagnostic", (event) => {
  console.log("Diagnostic event:", event.name, event.data);
  
  if (event.name === "videoStall") {
    if (event.data.stalled) {
      console.warn("Video stream stalled");
    } else {
      console.log(`Video recovered after ${event.data.durationMs}ms`);
    }
  }
});

Retry Strategies

Exponential Backoff

async function generateWithRetry(
  client: DecartClient,
  options: ProcessOptions,
  maxRetries = 3
) {
  let lastError;
  
  for (let attempt = 0; attempt < maxRetries; attempt++) {
    try {
      return await client.process(options);
    } catch (error) {
      lastError = error;
      
      // Don't retry on invalid input or API key
      if (typeof error === 'object' && error !== null && 'code' in error) {
        const sdkError = error as { code: string };
        if (
          sdkError.code === ERROR_CODES.INVALID_API_KEY ||
          sdkError.code === ERROR_CODES.INVALID_INPUT
        ) {
          throw error;
        }
      }
      
      // Exponential backoff: 1s, 2s, 4s
      const delay = Math.min(1000 * Math.pow(2, attempt), 10000);
      console.log(`Retry ${attempt + 1}/${maxRetries} after ${delay}ms`);
      await new Promise(resolve => setTimeout(resolve, delay));
    }
  }
  
  throw lastError;
}

// Usage
try {
  const blob = await generateWithRetry(client, {
    model: models.image("lucy-pro-t2i"),
    prompt: "a sunset",
  });
} catch (error) {
  console.error("Failed after retries:", error);
}

Queue Polling with Timeout

async function pollJobWithTimeout(
  client: DecartClient,
  jobId: string,
  timeoutMs = 300000 // 5 minutes
) {
  const startTime = Date.now();
  
  while (true) {
    if (Date.now() - startTime > timeoutMs) {
      throw new Error(`Job ${jobId} timed out after ${timeoutMs}ms`);
    }
    
    try {
      const status = await client.queue.status(jobId);
      
      if (status.status === "completed") {
        return await client.queue.result(jobId);
      }
      
      if (status.status === "failed") {
        throw new Error(`Job failed: ${status.error}`);
      }
      
      // Wait before polling again
      await new Promise(resolve => setTimeout(resolve, 2000));
    } catch (error) {
      console.error("Error polling job:", error);
      throw error;
    }
  }
}

// Usage
try {
  const job = await client.queue.submit({
    model: models.video("lucy-pro-t2v"),
    prompt: "flying bird",
  });
  
  const blob = await pollJobWithTimeout(client, job.job_id);
  console.log("Video ready");
} catch (error) {
  console.error("Video generation failed:", error);
}

React Error Boundaries

For React applications, use error boundaries to catch rendering errors:

import { Component, type ReactNode } from "react";

interface Props {
  children: ReactNode;
  fallback: ReactNode;
}

interface State {
  hasError: boolean;
  error?: Error;
}

class ErrorBoundary extends Component<Props, State> {
  constructor(props: Props) {
    super(props);
    this.state = { hasError: false };
  }

  static getDerivedStateFromError(error: Error) {
    return { hasError: true, error };
  }

  componentDidCatch(error: Error, errorInfo: any) {
    console.error("Error caught by boundary:", error, errorInfo);
  }

  render() {
    if (this.state.hasError) {
      return this.props.fallback;
    }

    return this.props.children;
  }
}

// Usage
function App() {
  return (
    <ErrorBoundary fallback={<div>Something went wrong</div>}>
      <VideoStream prompt="anime style" />
    </ErrorBoundary>
  );
}

Best Practices

Always handle errors - Never ignore try/catch blocks
Check error codes - Use specific error codes to provide better feedback
Retry transient errors - Network issues, timeouts, etc.
Don’t retry permanent errors - Invalid API keys, bad input, etc.
Log errors - Keep track of failures for debugging
Show user-friendly messages - Don’t expose technical details to users
Monitor connection state - React to state changes in real-time connections
Use diagnostic events - Track video stalls and other issues
Set timeouts - Don’t wait indefinitely for responses
Clean up resources - Disconnect clients, revoke object URLs, etc.

Next Steps

React Integration - React-specific patterns
Real-Time API Reference - Connection API details
Process API Reference - Synchronous generation API
Queue API Reference - Async job API

Getting Started

Core Concepts

Real-time API

Queue API

Process API

Guides

Error Types

Error Codes

Basic Error Handling

Process API (Synchronous)

Queue API (Async Jobs)

Real-Time Error Handling

Diagnostic Events

Retry Strategies

Exponential Backoff

Queue Polling with Timeout

React Error Boundaries

Best Practices

Next Steps

Build docs developers (and LLMs) love

Getting Started

Core Concepts

Real-time API

Queue API

Process API

Guides

​Error Types

​Error Codes

​Basic Error Handling

​Process API (Synchronous)

​Queue API (Async Jobs)

​Real-Time Error Handling

​Diagnostic Events

​Retry Strategies

​Exponential Backoff

​Queue Polling with Timeout

​React Error Boundaries

​Best Practices

​Next Steps

Build docs developers (and LLMs) love

Error Types

Error Codes

Basic Error Handling

Process API (Synchronous)

Queue API (Async Jobs)

Real-Time Error Handling

Diagnostic Events

Retry Strategies

Exponential Backoff

Queue Polling with Timeout

React Error Boundaries

Best Practices

Next Steps