Write, clean, format, transform and inspect text in one private workspace. Preview each step, apply it to your document, undo edits and compare against a fixed baseline without uploading your words.
Paste text or open a UTF-8 file up to 1 MiB. Document and applied output limit: 200,000 UTF-16 units. Line endings use LF.
Counts update locally. Word and sentence boundaries are approximate.
The first preview pins your starting document as a baseline. You can replace that baseline explicitly.
Preview does not change your document. Apply a step, then choose another to build a workflow. Reports cannot be applied as document edits.
Statistics follow the working document. Character counts use grapheme clusters when available; limits count UTF-16 units. Regex: 100,000 input units; encoding: 50,000 input units for encode; slugs: 20,000; Unicode inspection: 10,000; invisible-character operations: 100,000. Line operations: up to 20,000 lines. Diff: 100,000 units and 2,000 lines or word segments per side. Reports show up to 200 Unicode or invisible-character rows.
Need a dedicated view? AI Text Detector · Unicode Inspector · Text Diff Checker. Text Studio does not autosave documents.
Your text is never uploaded to any server; transformations, statistics and analysis run locally in disposable workers using shared self-hosted code. You can chain operations without moving text between separate websites. No account, remote API or automatic document storage is used. You control when a preview becomes an edit.
Word and character statistics, reading/speaking estimates, case conversion, whitespace cleanup, punctuation normalization, selected invisible-character cleanup, Unicode NFC/NFKC normalization, duplicate removal, sorting, line reversal and numbering, literal and regex replacements, Base64/Base64URL/URL/HTML encoding, slug creation, email/URL extraction, Lorem Ipsum, Unicode/invisible-character reports and baseline diff. Dedicated standalone tools remain available. AI Text Detector is linked separately and is not automatically run on your document.
Each operation reads the current working document and creates a preview. Apply replaces the document with that preview, allowing the next operation to use the result immediately. Reports are read-only and cannot be applied as edits. Undo and redo restore prior document states; manual typing is grouped into editing bursts. History is bounded to 20 snapshots and two million UTF-16 text units across undo and redo, so older snapshots can be released.
The first preview pins the current document as a baseline; importing a file or loading the example also sets one. It stays fixed while you edit or apply steps. Set comparison baseline explicitly replaces it. Compare reports changes from that baseline, and Restore baseline returns its text to the document as an undoable edit. Clear studio removes the baseline.
Document and applicable preview output: 200,000 UTF-16 units. Import: a valid UTF-8 file up to 1 MiB, still subject to the document limit. Line operations: 20,000 lines. Regex: 100,000 input units, 2,000 pattern units and 20,000 replacement units. Encoding input: 50,000 units for encode. Slugs: 20,000 units. Unicode inspection: 10,000; invisible-character analysis/cleanup: 100,000. Diff: 100,000 units and 2,000 lines or word segments per side. Unicode and invisible reports show at most 200 rows, explicitly marked. Inputs are not silently truncated.
Regex runs only in a disposable worker and is stopped after three seconds. Other operations and live statistics have five-second response limits. Editing the document, changing operation settings, Cancel and Clear terminate pending transformations and discard late results. A slow analysis can time out without changing your text.
UTF-8 byte counts describe the current text. Character counts use grapheme segmentation when available, with a code-point fallback. Words and sentences use browser-dependent language boundaries; estimates can differ by language and browser. Reading and speaking estimates use the selected words-per-minute rates, without extra speaking pauses. They are planning estimates, not measurements of an individual reader.
No. Operations are explicit and previewed because trimming indentation, changing punctuation, Unicode compatibility normalization or removing joiners and variation selectors can alter meaning, code or emoji. Invisible cleanup starts with no groups selected. Many line-oriented helpers normalize line endings to LF. Review each preview before applying, especially for source code or multilingual text.
No. Text, extracted candidates, decoded HTML entities and reports are displayed as plain text. No URL is visited, email contacted or decoded script executed. Base64 decoding expects valid UTF-8 text, not arbitrary binary files. Extraction is a practical candidate finder and does not verify that an address or destination exists.
No. Text Studio does not autosave documents or store input history in browser storage. Clear studio removes the current document, baseline, previews, find/replacement values and undo/redo snapshots, stops workers and revokes application-held download URLs. Leaving the page also cleans up. Download a TXT file if you want to keep your work.
