BeeMage — Explore image processing through visual pipelines
BeeMage is a client-side image-processing playground that lets you build, run, and inspect visual pipelines locally — without servers or cloud processing.
BeeMage focuses on exploration, understanding, and manual refinement of image-processing workflows, not black-box automation.
BeeMage is available in multiple delivery formats:
- Static web application (standalone, browser-based)
- Chrome Extension (Manifest V3, Side Panel)
- Android application (WebView-based wrapper)
Each delivery runs the same application core, adapted to its runtime environment.
No uploads.
No server-side processing.
Delivery projects live under apps/:
apps/demo(static web)apps/extension(Chrome extension)apps/android-web(Android web bundle)apps/android-native(Android app wrapper)
You can try BeeMage immediately in your browser:
The demo runs the same panel UI as the extension and requires no installation.
BeeMage focuses on one principle:
Making image-processing steps visible, inspectable, and understandable.
With BeeMage, you can:
- Load an image directly in the application
- Run explicit image-processing pipelines
- Observe intermediate representations (edges, masks, cleaned regions, SVG)
- Adjust parameters and see results immediately
- Execute pipelines step by step or end to end
- Export results as PNG or SVG
- Create, import, and manage custom pipelines visually (Builder tab)
The workflow is manual, visual, and iterative by design.
BeeMage deliberately avoids one-click automation in favor of understanding how processing decisions affect results.
BeeMage is built around a pipeline-based execution model.
A pipeline is a sequence of typed operations, for example:
image → edge detection → threshold → morphology → svg
Each step produces an explicit artifact (image, mask, or svg) that can be previewed and inspected.
Pipelines can be:
- selected from predefined examples
- built visually via drag & drop
- imported or exported as JSON
- extended with parameter presets (“recipes”)
The Android version embeds BeeMage as a local web application inside a native Android wrapper.
- The UI and processing logic are identical to the web version
- All image processing remains fully client-side
- No backend, no cloud services
Android is treated as a first-class runtime, not a separate codebase.
See:
apps/android-web/README.md— Android web bundle (web artifact embedded into the app)apps/android-native/README.md— Native wrapper (Gradle/Android Studio), APK/AAB build + publishing notes
The static web version is the simplest way to use BeeMage.
- Runs in any modern browser
- No installation required
- Identical UI and behavior to other deliveries
- Includes optional OpenCV (WebAssembly) support for experimentation
See:
apps/demo/README.md— Demo host project (Vite) and build/release scripts
This version is used for:
- the public demo
- documentation examples
- OpenCV comparison and exploration
👉 Live demo: https://nathabee.github.io/beemage/demo/
The Chrome Extension embeds BeeMage as a side panel application using Manifest V3.
- Same UI and pipeline system as the web version
- Integrated into the browser workflow
- Runs entirely client-side
- Uses only the native (TypeScript) execution engine
See:
apps/extension/README.md— Extension host project (esbuild), build/zip workflow, packaging notes
The extension is currently not yet published on the Chrome Web Store.
It is available for development builds and testing only.
The Chrome Extension deliberately disables OpenCV to ensure:
- deterministic behavior
- reduced bundle size
- full reviewability of execution
The standalone web version includes an engine strategy layer that allows selected operations to run using different implementations.
- Written in TypeScript
- Small, transparent, and inspectable
- Used in all delivery formats
- Emphasizes clarity and learning
- Available in the web demo
- Loaded locally (no CDN)
- Used selectively for comparison and experimentation
OpenCV support is currently partial and experimental and exists to explore:
- algorithmic differences
- parameter sensitivity
- trade-offs between simplicity and robustness
The Chrome Extension always runs using the native engine.
BeeMage is well suited for:
- Exploring image-processing concepts visually
- Learning and teaching classic image-processing techniques
- Extracting and refining outlines or silhouettes
- Comparing different processing strategies
- Prototyping pipelines without desktop software
- Preparing base assets for illustration or vector work
It is not intended to replace full-scale image editors or batch automation tools.
BeeMage is designed to be transparent and review-friendly.
- Runs entirely locally
- No analytics
- No tracking
- No advertising
- No uploads
Privacy policy: docs/privacy.html
Detailed documentation is available in the docs/ directory and on the project site:
👉 https://nathabee.github.io/beemage/
Key documents:
- Presentation and concepts:
docs/presentation.md - Architecture overview:
docs/architecture.md - Installation and deployment:
docs/installation.md - Versioning and development notes:
docs/version.md
BeeMage is in active development.
The core pipeline system is stable, while features and UI continue to evolve.
Feedback, testing, and issue reports are welcome.
- Homepage: https://nathabee.github.io/beemage/
- Live demo: https://nathabee.github.io/beemage/demo/
- Issues and support: https://github.com/nathabee/beemage/issues
BeeMage is licensed under the MIT License.
See the LICENSE file for the full text.
- Document updated for version:
0.2.8