moondream 1.2.2

Official Python client for Moondream, a fast and efficient vision language model.

Moondream Python Client Library

Official Python client library for Moondream, a fast multi-function VLM. This client can target Moondream Cloud or run locally via Photon — on NVIDIA GPUs (Linux x86_64 / aarch64 or Windows) or Apple Silicon Macs.

Capabilities

Moondream goes beyond the typical VLM "query" ability to include more visual functions:

Method	Description
caption	Generate descriptive captions for images
query	Ask questions about image content
detect	Find bounding boxes around objects in images
point	Identify the center location of specified objects
segment	Generate an SVG path segmentation mask for objects

Try it out on Moondream's playground.

Installation

pip install moondream

Quick Start

Choose how you want to run Moondream:

1. Moondream Cloud — Get an API key from the cloud console
2. Moondream Photon — High-performance local inference engine on NVIDIA GPUs (Linux / Windows) or Apple Silicon Macs (macOS 13+). Requires an API key.

import moondream as md
from PIL import Image

# Initialize with Moondream Cloud
model = md.vl(api_key="<your-api-key>")

# Or initialize with local inference (Photon — NVIDIA GPU or Apple Silicon)
model = md.vl(api_key="<your-api-key>", local=True)

# Load an image
image = Image.open("path/to/image.jpg")

# Generate a caption
caption = model.caption(image)["caption"]
print("Caption:", caption)

# Ask a question
answer = model.query(image, "What's in this image?")["answer"]
print("Answer:", answer)

# Stream the response
for chunk in model.caption(image, stream=True)["caption"]:
    print(chunk, end="", flush=True)

API Reference

Constructor

model = md.vl(api_key="<your-api-key>")                        # Cloud
model = md.vl(api_key="<your-api-key>", local=True)            # Photon (local: NVIDIA GPU or Apple Silicon)
model = md.vl(api_key="<your-api-key>", model="moondream3-preview/ft_id@step")  # Finetune

Methods

caption(image, length="normal", stream=False)

Generate a caption for an image.

Parameters:

• image — Image.Image or EncodedImage
• length — "normal", "short", or "long" (default: "normal")
• stream — bool (default: False)

Returns: CaptionOutput — {"caption": str | Generator}

caption = model.caption(image, length="short")["caption"]

# With streaming
for chunk in model.caption(image, stream=True)["caption"]:
    print(chunk, end="", flush=True)

---

query(image, question, stream=False)

Ask a question about an image.

Parameters:

• image — Image.Image or EncodedImage
• question — str
• stream — bool (default: False)

Returns: QueryOutput — {"answer": str | Generator}

answer = model.query(image, "What's in this image?")["answer"]

# With streaming
for chunk in model.query(image, "What's in this image?", stream=True)["answer"]:
    print(chunk, end="", flush=True)

---

detect(image, object)

Detect specific objects in an image.

Parameters:

• image — Image.Image or EncodedImage
• object — str

Returns: DetectOutput — {"objects": List[Region]}

objects = model.detect(image, "car")["objects"]

---

point(image, object)

Get coordinates of specific objects in an image.

Parameters:

• image — Image.Image or EncodedImage
• object — str

Returns: PointOutput — {"points": List[Point]}

points = model.point(image, "person")["points"]

---

segment(image, object, spatial_refs=None, stream=False)

Segment an object from an image and return an SVG path.

Parameters:

• image — Image.Image or EncodedImage
• object — str
• spatial_refs — List[[x, y] | [x1, y1, x2, y2]] — optional spatial hints (normalized 0-1)
• stream — bool (default: False)

Returns:

• Non-streaming: SegmentOutput — {"path": str, "bbox": Region}
• Streaming: Generator yielding update dicts

result = model.segment(image, "cat")
svg_path = result["path"]
bbox = result["bbox"]  # {"x_min": ..., "y_min": ..., "x_max": ..., "y_max": ...}

# With spatial hint (point)
result = model.segment(image, "cat", spatial_refs=[[0.5, 0.5]])

# With streaming
for update in model.segment(image, "cat", stream=True):
    if "bbox" in update and not update.get("completed"):
        print(f"Bbox: {update['bbox']}")  # Available in first message
    if "chunk" in update:
        print(update["chunk"], end="")  # Coarse path chunks
    if update.get("completed"):
        print(f"Final path: {update['path']}")  # Refined path
        print(f"Final bbox: {update['bbox']}")

---

encode_image(image)

Pre-encode an image for reuse across multiple calls.

Parameters:

• image — Image.Image or EncodedImage

Returns: Base64EncodedImage

encoded = model.encode_image(image)

Types

Type	Description
Image.Image	PIL Image object
EncodedImage	Base class for encoded images
Base64EncodedImage	Output of encode_image(), subtype of EncodedImage
Region	Bounding box with x_min, y_min, x_max, y_max
Point	Coordinates with x, y indicating object center
SpatialRef	[x, y] point or [x1, y1, x2, y2] bbox, normalized to [0, 1]

Links

• Website
• Playground
• GitHub