How browser-local file conversion works
Published and reviewed · Based on the current LocalShift implementation
A browser-local converter downloads its application code, reads a file you select on your device, and creates the result on that device. In this version of LocalShift, conversion input and output are not sent to a processing server.
Separate loading the website from converting a file
Your browser requests HTML, styles and scripts from the website host. The host necessarily receives connection information for those page requests. That is different from uploading the contents of a selected file.
When you choose a file, the page reads it using browser file APIs. Data conversion runs in a worker; image conversion uses browser decoding and encoding; PDF tools use a JavaScript PDF library. Processing speed depends on your device.
What happens to the result
The result is placed in browser memory and made available as a downloadable file. Selecting Download saves a copy through the browser. LocalShift does not add it to a cloud account.
Clear all resets the workspace. Refreshing or closing the tab clears the application's working state, but files you downloaded remain on your device. Your browser can cache website assets, and a managed device may have additional policies.
How to check network behavior
Open your browser's developer tools and the Network panel, then use a non-sensitive sample. Load the tool and perform a conversion. You should see website-resource requests, including a lazily loaded PDF module when needed, rather than a request carrying the sample file.
This is a useful check for a particular version, not a permanent guarantee about every website or browser extension. Our browser regression checks cover all 30 tools and flag non-GET requests or requests outside the tested site. They are not an independent security audit.
Understand the limits
Local processing still needs memory and CPU. Text input is limited to 5 MB and individual images to 20 MB and 24 megapixels. PDF workspaces accept up to 30 MB. Merge and split tools allow 500 output pages; the PDF preview, page-image export, organization, numbering and watermark tools accept up to 100 pages.
PDF and ZIP cancellation terminates their background workers. An image already encoding can finish before its result is discarded. For valuable documents, keep originals and inspect the output. The Privacy page describes the current site's handling of website requests.
Merge PDFs locally: a worked workflow
Use Merge PDF for unsigned, unencrypted documents. Add your files, use the arrows to place them in order, and convert. The pages within each file retain their order. Our downloadable two-page sample A followed by one-page sample B yields three pages: A1, A2 and B1.
Use Split PDF afterwards if you only need selected pages. A selection such as 3, 1-2 preserves that order. Choose one combined PDF, one PDF per page, or groups of pages. Encrypted PDFs are rejected, and signatures, bookmarks and interactive document features are not guaranteed to survive.