NVIDIA NIM Image Analysis icon

NVIDIA NIM Image Analysis

Analyze images with NVIDIA NIM Vision Language Models

NVIDIA NIM Image Analysis

AI-generated

Summary

Send an image to a chosen NVIDIA NIM Vision Language Model for analysis by providing base64, data URL, or image URL input along with a prompt, and receive the model's generated descriptive response.

Inputs

  • Model (required) — Choose the NVIDIA NIM Vision Language Model for image analysis, selected from a searchable list or by model ID (format: owner/model-name).
  • Image Data (required) — The image to analyze, provided as base64-encoded data, a full data URL, or a direct image URL (supports JPG, JPEG, PNG).
  • Prompt — Text prompt directing the model on how to analyze the image (default: 'Describe this image in detail.').
  • Additional Options — Optional parameters including Max Tokens (response length), Temperature (randomness/creativity), and Top P (response diversity).

Output shape

a list of items each containing the JSON response returned from the NVIDIA NIM API for the image analysis request

Each input item is processed individually; if an error occurs and 'Continue On Fail' is enabled, an error object containing the error message, item index, model used, and timestamp is returned rather than halting execution. The response includes the complete JSON from the '/chat/completions' API endpoint.

Links

Discussion