Originality Checking

What It Does

The Originality Checking system scans your content for matches against existing web pages. It identifies sentences that appear elsewhere on the internet, calculates an overall originality score, and flags potential plagiarism so you can address it before publishing. This protects your brand reputation and helps avoid SEO penalties for duplicate content.

Originality checking runs automatically on every article you generate. After content generation completes, ILLIXIS queues an originality scan in the background. Results appear on the article's Originality tab within 30-60 seconds. No configuration required.

How It Works

Fingerprint Selection

ILLIXIS selects "fingerprint" sentences from your article—phrases most likely to identify copying. The system prioritizes sentences that contain:

• Numbers and statistics: "Sales grew 47% in Q3 2025"
• Proper nouns: "According to the World Economic Forum..."
• Technical terms: Specialized vocabulary unique to the topic
• Named entities: People, organizations, locations, products

Web Search

Each fingerprint sentence is searched against the web using the SERP API. This checks whether the exact phrasing (or close variations) appears on other websites.

Similarity Analysis

For each search result, ILLIXIS calculates similarity using:

• N-gram matching (3-grams): Catches paraphrasing that simple word overlap misses
• Sequence matching: Finds partial copies within longer text
• Exact match detection: Identifies word-for-word copies

Scoring

The system produces:

• Originality score (0-100): Higher means more original. 90+ is excellent.
• Plagiarism percentage (0-100): The proportion of checked content that matches existing sources.
• Flagged sentences: Specific sentences with match details and source URLs.

Using the Originality Dashboard

Step 1: Access the Dashboard

Navigate to Content Hub > Originality from the main menu. The dashboard shows:

• Total Checked: Number of articles scanned
• Average Score: Mean originality score across all checks
• Flagged Count: Articles below your minimum score threshold
• Excellent Count: Articles scoring 90 or above

Step 2: Run a Check

There are two ways to check content:

Automatic (default): Originality checking runs automatically after every article is generated. Results appear on the article detail page.

Manual:

1. Find your content in the "Unchecked Content" section on the dashboard
2. Click the Check button next to the article
3. Wait for the analysis to complete (typically 15-30 seconds)

Bulk Check:

1. Click Bulk Check on the dashboard to queue all unchecked published content
2. Up to 50 articles are queued per batch
3. Click again if more remain

Step 3: Review Results

After a check completes:

1. Open any article in the Content Hub
2. Click the Originality tab
3. View the originality score and any flagged sentences
4. Each flagged sentence shows the matching URL, similarity percentage, and a snippet from the matching source

Understanding Results

Score Ratings

| Rating | Score | Meaning |
|--------|-------|---------|
| Excellent | 90-100 | Highly original content, no significant matches found |
| Good | 70-89 | Mostly original with minor matches (common phrases, quotes) |
| Concern | 50-69 | Significant matches found—review flagged sentences |
| Poor | Below 50 | Major duplication detected—rewrite or add attribution |

Match Types

• Exact match: Word-for-word copy of 5+ consecutive words
• High similarity: Close paraphrase with 80%+ n-gram overlap
• Partial match: Significant overlap in phrasing or structure

Common False Positives

Not every match indicates plagiarism. Common legitimate matches include:

• Industry terminology: Standard phrases in your field
• Quoted statistics: Numbers and data points from public sources
• Common expressions: Widely-used phrases that aren't unique to any source
• Your own published content: Matches against your own website (use excluded domains to filter these out)

Handling Flagged Content

Option 1: Rewrite Flagged Sections

1. Review each flagged sentence and its matching source
2. Rewrite the section in your own words
3. Re-run the originality check to verify improvement

Option 2: Add Attribution

If the matching text is a legitimate quote or citation:

1. Add proper attribution (e.g., "According to [Source]...")
2. The match remains but the content is properly cited

Option 3: Exclude Your Own Domains

If matches are against your own published content:

1. Go to Settings on the Originality dashboard
2. Add your domain(s) to the Excluded Domains list
3. Future checks will skip matches from those domains

Configuring Settings

Access settings from the dashboard by clicking Settings in the header.

Minimum Score Threshold

Default: 70

Articles scoring below this threshold are flagged for review. Lower this to be more lenient; raise it for stricter originality standards.

Sentences to Check

Default: 5-10 sentences

Controls how many fingerprint sentences are searched. More sentences means more thorough checking but slightly higher cost.

Excluded Domains

Add domains to exclude from matching (e.g., your own website). Comma-separated list.

Cost

Originality checking uses the SERP API. Typical cost: ~$0.01-0.02 per article (15 sentence searches). This cost is included in your plan—no additional charges.

Tips for Best Results

• Check before publishing: Originality results are most useful as a final quality gate
• Exclude your own domains: Avoid false positives from your own published content
• Review flagged sentences in context: A 30% match on a single sentence usually isn't concerning
• Use the bulk check feature: Scan your existing published content library for potential issues
• AI-generated content is usually original: Most AI content scores 85-95% since it generates new text rather than copying

Related Features

• Fact-Checking - Verify factual claims in your articles
• Content Grading - Overall quality scoring for your articles
• Article Generation - Creating AI-generated content
• Content Hub - Managing all your content in one place