SparkpondSparkpond · Image Tools
How local processing works
From source file to validated result, without an upload endpoint.
Browser → Worker → result
The File API reads the source and transfers its bytes to a dedicated Web Worker. The Worker detects, decodes, transforms, encodes, and verifies the output before the browser enables download.
Network resources and no upload route
Only same-origin static GET requests may be needed for HTML, CSS, JavaScript, translations, the Service Worker, and on-demand WASM codecs. The application has no document or image upload endpoint and sends no filename, image bytes, preview, metadata, or result.
Memory and optional local storage
Files, previews, and results stay in memory by default and disappear when the session is cleared or the tab closes. The optional 24-hour retention setting uses IndexedDB only after explicit opt-in, expires records automatically, and provides a Clear now action.
Offline behavior
After the app shell and a required codec have been cached, supported operations can run offline. A format whose codec has never been downloaded needs one successful online load first. User images and results are never added to Cache Storage.
Limits and output verification
Available memory, browser codec support, pixel count, and animation can limit an operation. Every candidate is redetected and decoded, then checked for its signature, dimensions, alpha policy, non-empty pixels, extension, and prohibited metadata. A failed candidate is not offered for download.