Moondream2

Class: Moondream2BlockV1

Source: inference.core.workflows.core_steps.models.foundation.moondream2.v1.Moondream2BlockV1

This workflow block runs Moondream2, a multimodal vision-language model. You can use this block to run zero-shot object detection.

Type identifier

Use the following identifier in step "type" field: roboflow_core/moondream2@v1to add the block as as step in your workflow.

Properties

Name	Type	Description	Refs
name	str	Enter a unique identifier for this step..	❌
prompt	str	Optional text prompt to provide additional context to Moondream2..	✅
model_version	str	The Moondream2 model to be used for inference..	✅

The Refs column marks possibility to parametrise the property with dynamic values available in workflow runtime. See Bindings for more info.

Available Connections

Compatible Blocks

Check what blocks you can connect to Moondream2 in version v1.

• inputs: QR Code Generator, Image Convert Grayscale, Google Gemini, Dynamic Crop, Blur Visualization, SIFT, Bounding Box Visualization, Stability AI Outpainting, Camera Focus, Slack Notification, Keypoint Visualization, Trace Visualization, Instance Segmentation Model, Polygon Visualization, Ellipse Visualization, Model Comparison Visualization, OpenAI, Anthropic Claude, Local File Sink, Triangle Visualization, Polygon Zone Visualization, Halo Visualization, LMM, Stability AI Image Generation, Florence-2 Model, Single-Label Classification Model, Circle Visualization, Email Notification, Google Vision OCR, Google Gemini, Clip Comparison, Camera Focus, Anthropic Claude, Object Detection Model, Instance Segmentation Model, Perspective Correction, CSV Formatter, Reference Path Visualization, Corner Visualization, Color Visualization, Twilio SMS/MMS Notification, VLM as Classifier, Multi-Label Classification Model, Image Slicer, OpenAI, Stitch OCR Detections, Camera Calibration, Image Blur, VLM as Detector, Dot Visualization, Roboflow Custom Metadata, Image Threshold, Model Monitoring Inference Aggregator, Morphological Transformation, Label Visualization, Background Color Visualization, Classification Label Visualization, OCR Model, Roboflow Dataset Upload, Keypoint Detection Model, Mask Visualization, Pixelate Visualization, Absolute Static Crop, Keypoint Detection Model, Webhook Sink, Grid Visualization, Contrast Equalization, Image Preprocessing, Google Gemini, Relative Static Crop, Stability AI Inpainting, Image Contours, Line Counter Visualization, Stitch Images, OpenAI, Crop Visualization, OpenAI, Llama 3.2 Vision, Icon Visualization, SIFT Comparison, Depth Estimation, Twilio SMS Notification, Single-Label Classification Model, Florence-2 Model, Background Subtraction, LMM For Classification, Multi-Label Classification Model, Object Detection Model, EasyOCR, CogVLM, Image Slicer, Roboflow Dataset Upload, Email Notification
• outputs: Corner Visualization, Color Visualization, Detections Combine, Detections Stitch, Stitch OCR Detections, Line Counter, Byte Tracker, Detections Transformation, Roboflow Custom Metadata, Dot Visualization, Model Monitoring Inference Aggregator, Velocity, Dynamic Crop, Blur Visualization, Detection Offset, Label Visualization, Background Color Visualization, Bounding Box Visualization, Time in Zone, Byte Tracker, Roboflow Dataset Upload, Trace Visualization, Path Deviation, Detections List Roll-Up, Pixelate Visualization, PTZ Tracking (ONVIF).md), Ellipse Visualization, Line Counter, Model Comparison Visualization, Detections Consensus, Size Measurement, Triangle Visualization, Distance Measurement, Byte Tracker, Crop Visualization, Time in Zone, Path Deviation, Florence-2 Model, Icon Visualization, Detections Stabilizer, Circle Visualization, Time in Zone, Florence-2 Model, Camera Focus, Detections Classes Replacement, Detections Filter, Detections Merge, Segment Anything 2 Model, Perspective Correction, Roboflow Dataset Upload, Overlap Filter

Input and Output Bindings

The available connections depend on its binding kinds. Check what binding kinds Moondream2 in version v1 has.

Bindings

• input

  • images (image): The image to infer on..
  • prompt (string): Optional text prompt to provide additional context to Moondream2..
  • model_version (roboflow_model_id): The Moondream2 model to be used for inference..

• output

  • predictions (object_detection_prediction): Prediction with detected bounding boxes in form of sv.Detections(...) object.

Example JSON definition of step Moondream2 in version v1

{
    "name": "<your_step_name_here>",
    "type": "roboflow_core/moondream2@v1",
    "images": "$inputs.image",
    "prompt": "my prompt",
    "model_version": "moondream2/moondream2_2b_jul24"
}