client - VLM Run

Client Configuration

Initialization

import { VlmRun } from "vlmrun";

const client = new VlmRun({
  apiKey: "your-api-key",
  // Optional configuration options
  baseURL: "https://api.vlm.run/v1",
});

Configuration Options

Option	Type	Description	Default
`apiKey`	`string`	Your VLM Run API key	Required
`baseURL`	`string`	Custom API base URL	`https://api.vlm.run/v1`
`timeout`	`number`	Request timeout in milliseconds	`120000`
`maxRetries`	`number`	Number of retries for failed requests	`5`

The gateway base URL is not a client option. Set VLMRUN_GATEWAY_URL in the environment to override it.

Client Components

// File operations
client.files.upload({
  /* ... */
});
client.files.get();

client.files.list();

// Model operations
client.models.list();

// Image predictions
client.image.generate({
  /* ... */
});

// Document predictions
client.document.generate({
  /* ... */
});

// Agent execution and OpenAI-compatible chat completions
client.agent.execute({
  /* ... */
});
client.agent.completions.create({
  /* ... */
});

// OpenAI-compatible model gateway (OCR / VLM models)
client.gateway.completions.create({
  /* ... */
});

// Skills, executions, and artifacts
client.skills.list();
client.executions.list();
client.artifacts.list({
  /* ... */
});

Error Handling

import { VlmRun, APIError, VLMRunError } from "vlmrun";

try {
  const client = new VlmRun({
    apiKey: "invalid-key",
  });
  await client.models.list();
} catch (error) {
  if (error instanceof APIError) {
    console.error("API Error:", error.message);
    console.error("Status:", error.http_status);
  } else if (error instanceof VLMRunError) {
    console.error("SDK Error:", error.message);
  } else {
    console.error("Unknown Error:", error);
  }
}

TypeScript Interfaces

class APIError extends VLMRunError {
  message: string;
  http_status?: number;
  headers?: Record<string, string>;
  request_id?: string;
  error_type?: string;
  suggestion?: string;
}

class VLMRunError extends Error {
  message: string;
}