Article Plagiarism Checker

Professional online text similarity checker supporting multiple algorithms to detect text duplication and similarity.

Bug report

Loading tool, please wait...

If the online tool fails to load or run, try the desktop tool.https://tools.yikeaigc.com/

Tool Usage

Return to old version

Smart Text Similarity Checker

Professional text similarity analysis tool based on multiple AI algorithms

Cosine Similarity Jaccard Index Longest Common Subsequence N-gram Matching Edit distance Semantic Similarity

AOriginal Text (Reference)

📄 Drag files here to upload

📁 Choose file to upload

BComparison Text (To Check)

📄 Drag files here to upload

📁 Choose file to upload

🔧 Detection Settings

Similarity Threshold (%)

N-gram Length

Minimum Match Length

Ignore Punctuation

Ignore case

Language Mode

Select Detection Algorithm

Cosine Similarity

Similarity calculation based on vector space model

Jaccard Index

Similarity metric based on set intersection over union

Longest Common Subsequence

Dynamic programming algorithm for sequence similarity

N-gram Matching

Matching algorithm based on character/word fragments

Edit distance

Distance metric for string edit operations

Semantic Similarity

Semantic understanding algorithm based on word vectors

Analyzing text similarity...

Instructions

Software Usage Instructions

Enter text content: Enter the reference text in the "Original Text" box on the left, and enter the text to be checked in the "Comparison Text" box on the right. You can also upload a txt text file by clicking the "Select File to Upload" button.
Select a detection algorithm: In the parameter settings panel, select the similarity detection algorithm you want to use:
- Cosine Similarity: Similarity calculation based on the vector space model
- Jaccard Coefficient: A similarity measure based on the intersection-over-union of sets
- Longest Common Subsequence: Uses a dynamic programming algorithm to calculate sequence similarity
- N-gram Matching: A matching algorithm based on character/word segments
- Edit Distance: A distance measure for string edit operations
- Semantic Similarity: A semantic understanding algorithm based on word vectors
Adjust detection parameters:
- Similarity Threshold: Set the percentage threshold for similarity judgment (0-100%)
- N-gram Length: Set the character segment length for the N-gram algorithm
- Minimum Match Length: Set the minimum number of matching characters
- Ignore Options: Choose whether to ignore punctuation and case
- Language Mode: Select Chinese, English, or Chinese-English mixed mode
Batch file processing: Supports uploading multiple txt files simultaneously for batch detection. The interface displays the first 20 files, but all uploaded files will be processed.
Start detection: Click the "Start Detection" button, and the system will use the selected algorithm to analyze text similarity.
View results: After detection is complete, view the detailed similarity report, including detection results from each algorithm and visual charts.
Export results: Click the "Export Results" button to package and download the detection report and raw data as a ZIP file.

FAQ

A: Each algorithm uses a different calculation principle. Cosine Similarity focuses on word frequency distribution, the Jaccard Coefficient focuses on vocabulary overlap, the LCS algorithm focuses on character sequences, N-gram focuses on segment matching, Edit Distance focuses on the cost of character modifications, and Semantic Similarity focuses on semantic understanding. It is recommended to make a judgment by combining the results of multiple algorithms.

A: Recommended threshold settings: 70% or above indicates high similarity, 50-70% indicates moderate similarity, 30-50% indicates low similarity, and below 30% indicates basically not similar. For academic checks, a setting of 60-70% is recommended; for content originality checks, a setting of 40-50% is recommended.

A: N-gram length affects detection accuracy: length 2-3 is suitable for detecting character-level similarity, length 4-5 is suitable for detecting word-level similarity, and length 6-8 is suitable for detecting phrase-level similarity. For Chinese text, 3-4 is recommended; for English text, 4-5 is recommended.

A: No. All text similarity checks are performed locally on your device. Text content and files will not be uploaded to the server, fully protecting your data privacy and security.

A: Currently, plain text files in TXT format are supported. It is recommended to first convert documents in formats such as Word or PDF to TXT format, or directly copy the text content into the input box for detection.

Answer: Text similarity detection consumes computing resources. Recommendations: 1) reduce the number of algorithms used simultaneously; 2) split long texts into sections for detection; 3) keep the text length for a single detection within 100,000 characters; 4) close other programs that are using memory.

Answer: The exported ZIP file contains: Similarity Detection Report.txt (a human-readable detailed report) and detection_data.json (machine-readable structured data). The report includes information such as the similarity percentages for each algorithm, the average value, and detailed descriptions.