Plagiarism Checker
Check your text for plagiarism in seconds — free, no account required, and nothing leaves your device. The AI model runs entirely in your browser and compares sentence-by-sentence against the reference text you provide.
This tool compares your text against the reference text you paste — it does not search the internet. Up to 30 sentences per text area are analysed.
How It Works
Paste your text
Drop in any writing you want to check — essays, reports, or any text up to 5,000 characters. Nothing is sent anywhere.
Add a reference
Paste the source or reference material you want to compare against — a book excerpt, article, or original source.
AI encodes locally
On first use, your browser downloads the AI model (~23 MB) once and saves it locally. The MiniLM model then encodes every sentence into a semantic embedding right here in your browser.
Review colour-coded results
Each sentence is rated: red for flagged (≥ 85% similar), yellow for suspicious (70–84%), and green for original (< 70%). The best-matching reference sentence is shown for flagged and suspicious lines.
How to Use the Plagiarism Checker
- Paste or type the text you want to check into the Your Text area — up to 5,000 characters.
- Paste the source or reference material into the Reference Text area.
- Click Check for Plagiarism. On first use, a one-time model download (~23 MB) is required — your browser saves it locally for instant reuse.
- Review the overall similarity score and the colour-coded sentence breakdown. Flagged sentences (≥ 85% similarity) show the best-matching reference sentence alongside them.
- Click Copy Results to copy the full report to your clipboard, or Reset to start a new check.
Features
- Your text stays private: The AI model runs entirely in your browser — nothing is uploaded, stored, or shared.
- No account required: Open the page and start checking straight away.
- Sentence-level analysis: Each sentence in your text is individually scored and colour-coded for instant readability.
- Best-match display: For flagged and suspicious sentences, the tool shows the most similar sentence from your reference text so you can see exactly what triggered the flag.
- Works offline: After the one-time model download, the tool runs with no internet connection.
- Compares against your reference: The tool checks your text against source material you provide — your text is never sent to an external search engine or database.
- Powered by MiniLM-L6-v2: A lightweight sentence transformer by Sentence Transformers that produces 384-dimensional semantic embeddings. One-time download of ~23 MB.
Frequently Asked Questions
Is my text uploaded to your servers?
No. Your text never leaves your device. The entire process runs inside your browser — nothing is sent to or stored on any server.
Does this tool check against the internet?
No. This tool compares your submitted text against a reference text that you paste yourself. It does not search the web or any external database. This is by design — it means your text stays completely private and the tool works offline.
Is this free to use?
Yes, completely free. No account required, no usage limits, and no hidden charges.
Why is the first check slow?
The first time you use the tool, your browser downloads the AI model (~23 MB). After that one-time download, the model loads instantly from local storage and the tool works offline too.
How does the similarity score work?
Each sentence is converted into a 384-dimensional semantic embedding using the MiniLM model. Cosine similarity is then computed between every submitted sentence and every reference sentence. The highest score per submitted sentence determines its label: ≥ 85% = Flagged, 70–84% = Suspicious, < 70% = Original.
Does it work offline?
Yes. After the model is downloaded on your first use, it is saved to your browser's local storage. You can then run plagiarism checks with no internet connection.
Which AI model does this use?
all-MiniLM-L6-v2 by Sentence Transformers — a lightweight sentence embedding model that maps text into a 384-dimensional semantic space. It runs as a quantized ONNX model via Transformers.js in a Web Worker in your browser. The one-time download is ~23 MB.