streamlit-video-annotator 0.1.1

A Streamlit component for annotating videos with drawable regions and time-range markers

Streamlit Video Annotator

A custom Streamlit component for annotating videos with drawable regions and time-range markers. Perfect for computer vision, machine learning, and video analysis applications.

Features

• Video Playback Controls: Play, pause, and scrub through videos
• Multiple Drawing Tools:
  • Rectangle
  • Circle
  • Freedraw (path)
  • Arrow
• Time-Range Annotations: Mark start and end times for each annotation
• Color Customization: Choose from multiple colors for annotations
• Comments: Add text descriptions to annotations
• Annotation Management: View, edit, and delete existing annotations
• Internationalization: Customize all UI labels for different languages
• Responsive Design: Adjustable height and responsive layout

Installation

pip install streamlit-video-annotator

Try the Demo

Want to see it in action first? Clone the repository and run the demo app:

git clone https://github.com/mikaelnystrom/video_annotator.git
cd video_annotator
pip install -e .
streamlit run app.py

The demo app includes:

• Sample video (Big Buck Bunny)
• Multiple language options (English, Swedish, Spanish, French, German)
• Customizable colors and height
• Live annotation preview and data display

Quick Start

import streamlit as st
from streamlit_video_annotator import video_annotator

# Basic usage
result = video_annotator(
    video_url="https://example.com/video.mp4",
    height=600
)

# Handle new annotations
if result and result.get("newAnnotation"):
    st.write("New annotation created:", result["newAnnotation"])
    # Save to database, process, etc.

# Handle deletions
if result and result.get("deletedAnnotationId"):
    st.write("Annotation deleted:", result["deletedAnnotationId"])

Advanced Usage

With Existing Annotations

from streamlit_video_annotator import video_annotator

# Load existing annotations from your database
existing_annotations = [
    {
        "id": "annotation-1",
        "startTime": 5.0,
        "endTime": 10.0,
        "shape": {
            "type": "rectangle",
            "x": 0.2,
            "y": 0.3,
            "width": 0.4,
            "height": 0.3,
            "color": "#ff0000"
        },
        "comment": "Object of interest",
        "createdAt": "2026-01-08T10:00:00Z"
    }
]

result = video_annotator(
    video_url="https://example.com/video.mp4",
    existing_annotations=existing_annotations,
    height=700
)

Custom Labels (Internationalization)

# Swedish labels example
swedish_labels = {
    "play": "Spela",
    "pause": "Pausa",
    "tools": "Verktyg",
    "rectangle": "Rektangel",
    "circle": "Cirkel",
    "freedraw": "Frihand",
    "arrow": "Pil",
    "color": "Färg",
    "markStart": "Markera Start",
    "markEnd": "Markera Slut",
    "saveAnnotation": "Spara Annotering",
    "cancel": "Avbryt",
    "delete": "Ta bort"
}

result = video_annotator(
    video_url="video.mp4",
    labels=swedish_labels
)

Custom Colors

result = video_annotator(
    video_url="video.mp4",
    colors=["#FF5733", "#33FF57", "#3357FF", "#F033FF"]
)

API Reference

video_annotator()

video_annotator(
    video_url: str,
    existing_annotations: Optional[List[AnnotationData]] = None,
    height: int = 600,
    labels: Optional[Dict[str, str]] = None,
    colors: Optional[List[str]] = None,
    key: Optional[str] = None,
) -> Optional[Dict[str, Any]]

Parameters:

• video_url (str): Direct URL to the video file. Supports MP4, WebM, and other browser-compatible formats. Note: YouTube URLs are not supported.
• existing_annotations (list, optional): List of annotation dictionaries to display.
• height (int, optional): Component height in pixels. Default: 600.
• labels (dict, optional): Custom UI labels for internationalization.
• colors (list, optional): List of color hex codes for annotations. Default: ['#00ff00', '#ff0000', '#0000ff', '#ffff00', '#ff00ff', '#00ffff'].
• key (str, optional): Unique key for the component instance.

Returns:

Dictionary with:

• annotations: Full list of current annotations
• newAnnotation: Most recently added annotation (if any)
• deletedAnnotationId: ID of deleted annotation (if any)

Returns None if no changes occurred.

Data Structures

AnnotationData

{
    "id": str,              # Unique identifier
    "startTime": float,     # Start time in seconds
    "endTime": float,       # End time in seconds
    "shape": ShapeData,     # Shape information
    "comment": str,         # User comment
    "createdAt": str        # ISO 8601 timestamp
}

ShapeData

{
    "id": str,
    "type": str,           # 'rectangle', 'circle', 'path', or 'arrow'
    "color": str,          # CSS color (e.g., '#ff0000')

    # For rectangles:
    "x": float,            # 0-1 normalized
    "y": float,            # 0-1 normalized
    "width": float,        # 0-1 normalized
    "height": float,       # 0-1 normalized

    # For circles:
    "x": float,            # Center X (0-1 normalized)
    "y": float,            # Center Y (0-1 normalized)
    "radius": float,       # 0-1 normalized

    # For arrows:
    "x": float,            # Start X
    "y": float,            # Start Y
    "endX": float,         # End X
    "endY": float,         # End Y

    # For paths (freedraw):
    "points": [            # List of points
        {"x": float, "y": float},
        ...
    ]
}

Use Cases

• Computer Vision Training: Create labeled datasets for object detection and tracking
• Video Analysis: Mark regions of interest in research videos
• Quality Assurance: Annotate defects or issues in video footage
• Sports Analysis: Mark player positions and movements
• Medical Imaging: Annotate regions in medical video footage
• Educational Content: Create interactive video lessons with annotations

Requirements

• Python >= 3.8
• Streamlit >= 1.0.0

Development

Local Development Setup

1. Clone the repository
2. Install dependencies:

   pip install -e .
   cd video_annotator/frontend
   npm install
   

3. For frontend development, start the dev server and set the dev mode flag:

   npm start
   STREAMLIT_COMPONENT_DEV=true streamlit run app.py
   

Building for Production

cd video_annotator/frontend
npm run build

Publishing to PyPI

This package uses GitHub Actions with PyPI trusted publishing for secure, automated releases.

One-Time Setup

1. Configure PyPI Trusted Publisher:

   • Go to https://pypi.org/manage/account/publishing/
   • Add a new publisher with these details:
     • PyPI Project Name: streamlit-video-annotator
     • Owner: Your GitHub username (e.g., mikaelnystroms)
     • Repository name: video_annotator
     • Workflow name: publish.yml
     • Environment name: pypi

2. Create GitHub Environment:

   • Go to your repo Settings → Environments
   • Create a new environment named pypi
   • (Optional) Add protection rules like requiring reviewers

Publishing a Release

1. Update version in pyproject.toml and video_annotator/video_annotator.py

2. Commit and tag:

   git add .
   git commit -m "Release v0.1.1"
   git tag v0.1.1
   git push origin main --tags
   

3. Automated publish: The GitHub Action will automatically:

   • Build the package with uv
   • Publish to PyPI using trusted publishing
   • No API tokens needed!

Manual Publish (Alternative)

If you prefer to publish manually:

uv build
uv publish

License

MIT License - see LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Support

If you encounter any issues or have questions, please file an issue on GitHub.

Links

• GitHub Repository
• Streamlit Documentation
• Report Issues