A browser workbench for everyday files
Useful file tools, with the processing boundary in plain sight.
LocalUtils helps you prepare PDFs, inspect data, convert images, recognize text, create QR codes, and verify files without first sending the selected file to a conversion server. Each tool explains what runs locally, what it stores in your browser, and where its limits begin.
device → browser engine → result
Local processing is a technical boundary, not a slogan
A browser application still uses the network to download its interface, libraries, fonts, and optional language assets. Account authentication, feedback, billing, and advertising are also network-connected features. The important distinction is the file-processing path: the selected document, image, or data file is passed to code running in the browser instead of being posted to a LocalUtils conversion endpoint.
Different tools use different browser capabilities. PDF.js renders PDF previews, PDF-lib rebuilds PDF documents, piexifjs reads and changes JPEG metadata, heic2any decodes HEIC images, Papa Parse handles CSV data, Tesseract.js performs OCR, and dedicated workers keep several long-running tasks away from the main interface. A tool page names the relevant library and does not claim that every operation uses the same technology.
Local processing reduces one category of data transfer, but it does not make an unsafe device safe. Browser extensions, malware, shared profiles, unreviewed output, and files that contain unexpected structures remain practical risks. Treat every downloaded result as something to inspect before sharing or relying on it.
How to check the boundary yourself
- 01
Open the browser developer tools, select the Network panel, and clear the existing request list before choosing a non-sensitive test file.
- 02
Run the operation and look for an upload request containing the file. Normal asset, account, advertising, or OCR language-download requests are separate from a file-processing upload.
- 03
After required assets are cached, disconnect the network and repeat the operation with a test file. Expect sign-in, advertising, billing, and first-time downloads to stop working while supported local operations continue.
Choose the job, then read its limits
The directory is organized by the outcome you need. Every card links to an interactive tool and a page-specific explanation of supported files, output behavior, common mistakes, and a verification workflow.
PDF Utilities
Merge, split, rotate, and clean PDF metadata
The PDF toolbox is intended for document preparation rather than page-level editing. It can arrange several PDFs into one file, rotate every page in a selected document, divide one document into complete page ranges, and clear common document-information fields.
Limits to know before you start: Password-protected or damaged PDFs may fail to open.
Open the tool →02Mark and rasterize sensitive areas in a PDF
Drawing a black rectangle over text is not necessarily redaction. In many PDF editors, the original text remains under the rectangle and can still be copied, searched, or extracted.
Limits to know before you start: The tool cannot know whether every sensitive area was marked.
Open the tool →Image Processing
Inspect and remove selected JPEG metadata before sharing
A photo can reveal more than its visible pixels. JPEG metadata may contain camera details, capture time, software history, comments, orientation, and—when recorded—location coordinates.
Limits to know before you start: Visible information remains visible even after metadata is removed.
Open the tool →04Convert HEIC images to browser-friendly JPEG or PNG
HEIC stores images efficiently but is not accepted by every website, editor, or operating system. Conversion is useful when compatibility matters more than preserving the original container.
Limits to know before you start: Some HEIC variants, sequences, auxiliary images, or color profiles may not decode as expected.
Open the tool →05Extract editable English text from a clear image
Optical character recognition estimates characters from pixels; it does not retrieve an original text layer. Accuracy depends heavily on focus, contrast, orientation, typography, layout, and language.
Limits to know before you start: Handwriting, curved text, tables, multi-column layouts, decorative fonts, and low-resolution screenshots reduce accuracy.
Open the tool →06Create a static QR code and test it before publishing
A QR code is a visual encoding of the exact text entered into the generator. A URL code does not pass through a LocalUtils redirect and does not gain analytics, expiry, or destination editing.
Limits to know before you start: Long payloads create dense codes that need more pixels and physical space.
Open the tool →Data & Logs
Validate, format, minify, and scrub JSON deliberately
JSON is strict about quotes, commas, brackets, and value types. Formatting can make a document readable, while minification removes insignificant whitespace. Neither operation changes whether the data is sensitive.
Limits to know before you start: A valid document can still have the wrong schema or misleading values.
Open the tool →08Inspect large delimited files without loading every row at once
CSV is a family of conventions rather than one perfectly uniform format. Delimiters, quoting, embedded newlines, encodings, and uneven row lengths all affect how a file appears.
Limits to know before you start: Encoding detection is limited; a file may need conversion to UTF-8.
Open the tool →Security & Encoding
Calculate file checksums and compare them exactly
A checksum is a repeatable fingerprint of file bytes. Matching a value published by a trusted source can show that your downloaded bytes match that source's reference.
Limits to know before you start: A matching digest cannot detect malware when the trusted reference describes the same malicious file.
Open the tool →10Convert text between UTF-8, Base64, and hexadecimal
Base64 and hexadecimal are reversible text representations of bytes. They help move data through systems that expect printable characters, but they do not provide secrecy, authenticity, or access control.
Limits to know before you start: Arbitrary binary data may not decode to meaningful UTF-8 text.
Open the tool →Field notes for safer file handling
The guide library goes beyond button instructions. It explains why PDF overlays can fail, what photo metadata removal cannot hide, when a checksum proves identity but not safety, how OCR accuracy changes with the source image, and why Base64 is an encoding rather than protection.
A review-first workflow for merging, splitting, rotating, and cleaning PDFs
Prepare PDFs in the browser while preserving page order, accounting for every split page, and verifying metadata and output behavior.
Read the full guide →Mark and rasterize sensitive areas in a PDFHow to verify destructive PDF redaction instead of trusting a black box
Learn why PDF overlays can leave text recoverable, how LocalUtils rasterizes marked pages, and how to inspect the exported document.
Read the full guide →Inspect and remove selected JPEG metadata before sharingRemoving photo metadata is useful—then the pixel review begins
Inspect JPEG metadata, remove selected EXIF groups, and review the visible clues and unsupported data that can remain.
Read the full guide →Convert HEIC images to browser-friendly JPEG or PNGHEIC to JPEG or PNG: choose compatibility without pretending nothing changes
Understand browser HEIC conversion, output tradeoffs, metadata limits, memory use, and a practical comparison workflow.
Read the full guide →Extract editable English text from a clear imageBetter OCR starts before recognition and finishes with human review
Prepare source images for English Tesseract OCR, understand common errors, and verify names, numbers, layout, and punctuation.
Read the full guide →Create a static QR code and test it before publishingA QR code is only finished after the final artwork scans reliably
Create static QR codes with appropriate payload length, contrast, quiet space, output size, and destination verification.
Read the full guide →Validate, format, minify, and scrub JSON deliberatelyReadable JSON is not automatically valid for the job—or safe to share
Validate JSON syntax, understand serialization, apply field-name scrubbing carefully, and search the result for remaining sensitive values.
Read the full guide →Inspect large delimited files without loading every row at onceLarge CSV review is about parsing assumptions, not only row count
Inspect delimiters, quoting, encodings, bookmarks, row ranges, edits, and exports when a delimited file is too large for a spreadsheet.
Read the full guide →Calculate file checksums and compare them exactlyA checksum comparison is strong only when the reference is trustworthy
Calculate file digests locally, choose the publisher's algorithm, compare complete values, and understand what a match cannot prove.
Read the full guide →Convert text between UTF-8, Base64, and hexadecimalBase64 and hexadecimal describe bytes; they do not protect them
Convert UTF-8 text, Base64, and hexadecimal while accounting for byte encodings, padding, URL-safe variants, and unsafe decoded content.
Read the full guide →Before using a browser utility
Do the tools work without an internet connection?
Many processing operations can run after the page and required assets are loaded. First-time OCR language assets, authentication, feedback, billing, advertising, and application updates still require a network connection.
Are files saved anywhere?
Most tools keep working data in memory. The PDF toolbox keeps its queue in IndexedDB so it can survive a refresh, and the JSON formatter stores its current input and scrub settings in localStorage. Clear the tool and browser storage on a shared device.
Does local processing guarantee privacy?
No. It avoids uploading selected files to a LocalUtils processing endpoint, but it cannot protect a compromised device, a malicious extension, an unsafe browser profile, or information that remains visible in the result.
Can I rely on a result without reviewing it?
No. Conversions, OCR, redaction, parsing, and metadata changes can all encounter unusual files. Open the downloaded output in another viewer and check the pages, text, metadata, or checksum relevant to the job.
How can I report a problem?
Sign in and use Feedback in the sidebar. Describe the browser, the tool, the file type, and the steps that caused the problem, but do not attach or paste confidential file contents.