ReworkMyText automatically changes wording and selected sentence structures with deterministic rules. Paste text or upload a document, choose how much to rework, and compare the result side by side.
Runs in your browserNo LLMNo model downloadProtected facts checked
Choose a document or drop it herePDF, DOCX, DOCM, HTML, HTM, TXT, MD, CSV, JSON, XML · up to 25 MB Text is extracted in your browser. The rewritten result is plain text.
Light makes a small number of conservative wording/structure changes; Standard goes further; Deep uses the broadest deterministic rewrite rules. If no safe change can be made, the result will say No rewrite rather than treating unchanged text as a successful rework.
Original / extracted text0 characters
Browser mode does not send this text to ScrubMyText.
Rewritten0 characters
Ready
0%estimated wording change
0%sentences structurally transformed
0deterministic transformations
—protected-item preservation check
Protected items: numbers, URLs, email addresses, quotations, acronyms, code, and many multi-word proper names are masked before rewriting and restored afterward.
How ReworkMyText works without an LLM
Most paraphrasing products send text to a language model. ReworkMyText takes a different approach. It first normalizes copy/paste artifacts and temporarily protects values that should not be casually changed. It then applies a curated library of high-confidence phrase substitutions and conservative sentence transformations. Finally, it restores the protected values and checks that they survived exactly.
Light
Conservative wording cleanup. Best when you want minimal movement and easy review.
Standard
More phrase changes plus limited sentence restructuring. The default for everyday prose.
Deep
Uses the broadest rule set and more structural transformations. Best for text that can tolerate heavier editing.
What it tries to preserve
Numbers, percentages, dates and currency-like values
URLs and email addresses
Quoted passages and inline/code blocks
Acronyms and many multi-word proper names
Paragraph order and overall topic sequence
What document upload does
Uploads are parsed locally into text. PDF pages are extracted from the text layer; Word files are read from their document XML; text and web formats are read directly. ReworkMyText does not rebuild the original PDF or Word formatting—the output is editable plain text.
Scanned/image-only PDFs require OCR and are not supported in this version.
Why it may leave some text alone
A deterministic rewriter has less semantic understanding than a language model. When no high-confidence rule matches, ReworkMyText prefers to leave a passage unchanged instead of inventing a risky synonym or altering meaning. That makes the result more predictable, but also means some prose will change much more than other prose.
Use the change metrics correctly
The percentages describe differences between the original and rewritten text. They are not an AI-authorship score, watermark result, plagiarism score, or detector-evasion guarantee. ReworkMyText does not have access to secret provenance keys or third-party detector internals.
Developer API
The same deterministic rewrite engine can be exposed through ScrubMyText’s API. The REST endpoint accepts text plus a rewrite mode and returns the reworked text, transformation counts, preservation checks and change metrics. The API version is intended for automation and uses the existing ScrubMyText paid API-key system.
curl -X POST https://api.scrubmytext.com/v1/rework \
-H "Authorization: Bearer smt_live_YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{"text":"Your text here","mode":"standard"}'
The website version runs locally. The REST API necessarily sends the supplied API text to ScrubMyText for processing.
Frequently asked questions
Does ReworkMyText use ChatGPT, Claude, Gemini or another LLM?
No. The rewrite engine is a deterministic JavaScript rule system. It does not call a language-model API and does not download a local model.
Can I upload a PDF or Word document?
Yes. The browser can extract text from text-layer PDFs and DOCX/DOCM files, as well as common text and web formats. The result is rewritten plain text rather than a reconstructed copy of the original document.
Does my uploaded document leave my browser?
Not in the browser tool. File contents are parsed locally. Some open-source parsing library code may be loaded by the page, but the selected document is not intentionally uploaded to those library hosts.
Will every sentence be changed?
No. The engine changes text only when a rule is considered safe enough. Deep mode applies more rules, but deterministic rewriting cannot reliably paraphrase every possible sentence.
Does this prove text is human-written or remove a watermark?
No. The tool reports only the transformations it made. It does not certify authorship, provenance, watermark status, or the outcome of any third-party detector.
Why offer an API if the browser version is free?
The browser version is designed for people editing text interactively. The API is for software and agents that need the same deterministic behavior programmatically and at repeatable scale.
scrubmytext