One local workspace
The architecture is designed for future pipelines: extract, clean, deduplicate, compare and export without sending the working set to the backend.

Compare, clean, calculate and validate in one privacy-first workspace. Core operations run locally, pages explain their exact rules, and no tool pretends a thin form is a complete product.
Compare two columns and separate left-only, right-only and matching lines without uploading either list.
Open tool →2.Deduplicate lines while preserving first-seen order, with optional case and whitespace normalization.
Open tool →3.Sort text lines alphabetically, naturally or in reverse and keep the entire operation in your browser.
Open tool →4.Turn delimited text into lines or join lines with a custom separator in one privacy-first tool.
Open tool →5.Replace literal text or a bounded regular expression and preview the transformed result instantly.
Open tool →6.Extract web addresses, integers or decimal numbers from pasted text without a server request.
Open tool →7.Add consistent text before or after every non-empty line and export the transformed list.
Open tool →8.Measure lines, words, characters, unique values and the sum of detected numbers locally.
Open tool →9.Reveal zero-width, non-breaking, directional and other invisible Unicode characters that can break imports, URLs and SEO data.
Open tool →10.Find visually similar lines after safe Unicode normalization without sending a keyword, title or translation list to a server.
Open tool →Estimate file size from bitrate and duration, or solve for the bitrate that fits a target size.
Open calculator →02.Estimate transfer time with correct bit/byte units and optional network overhead.
Open calculator →03.Scale dimensions for 16:9, 9:16, 1:1 and custom media without distortion.
Open calculator →04.Remove tracking parameters while preserving the destination and meaningful media identifiers.
Open cleaner →The architecture is designed for future pipelines: extract, clean, deduplicate, compare and export without sending the working set to the backend.
Each indexable page represents a working task with its own explanation and examples. Query-state pages remain canonical to the task instead of multiplying thin URLs.
No arbitrary server-side URL fetching, no open relay, and no silent storage of pasted lists. Sharing a recipe will never include the source data.