Topical Drift · Beta

Start your first topical drift scan in minutes

Upload a sitemap or URL list, cluster pages by topic, and surface URLs that have drifted from their original intent before they quietly erode rankings and internal link signals.

Get Started – Run a Free Scan

No credit card • Read-only inputs (sitemap/URL list, GSC) • You own all exports

Free during beta Limits: Up to 500 pages per site and unlimited scans (fair use) while we scale infrastructure.

You’re ready if you can:

Grab your sitemap.xml (or a URL list)

Export GSC (pages + queries) Recommended

You can run a scan without GSC — but adding it helps us prioritize fixes by real demand (clicks, impressions, queries).

What a single scan gives you

A topical map: clusters built from your sitemap (or URL list) and the content we extract.
Per-URL drift scores that highlight pages pulling away from their core topic.
Radial + cluster visualizations to see how hubs, spokes, and outliers connect.
A prioritized checklist of fixes grouped by cluster and opportunity size.

Your data stays private. Files are used only to generate your drift report and exports.

How the scan works

Upload a sitemap (or URL list), and Topical Drift does the rest—content extraction, clustering, drift scoring, and reporting.

Add a Google Search Console export if you have it. It helps us prioritize fixes by real demand (clicks, impressions, and queries).

Upload your pages

Start with a sitemap XML or a URL list. Optionally include a GSC CSV export to connect results to performance.

We extract & embed

We crawl pages, extract the main content, and generate embeddings automatically— no embedding files required.

Cluster by topic

Pages are grouped into semantic clusters so you can see your site’s topical layout and where overlap or fragmentation is happening.

Review the map & action plan

Explore the radial map, inspect drifting URLs, and export a prioritized checklist for content, intent alignment, and internal linking fixes.

Go to upload step

What you’ll need to get started

You only need a sitemap or a URL list. Upload that and we’ll crawl, extract content, and build the scan automatically.

GSC data is recommended (not required) so we can tie drift to real demand—queries, impressions, clicks, CTR, and position.

Sitemap XML Required (choose one)

Use sitemap.xml to generate your page list and start the scan.

URL list (TXT/CSV) Required (choose one)

No sitemap? Upload a simple list of URLs—one per line (or a CSV with a URL column).

Google Search Console export (CSV) Recommended

Adds demand + performance context (queries, impressions, clicks, CTR, position) so your drift report prioritizes what matters most.

Don’t worry about embeddings or internal link exports — we generate those during the scan.

Continue to Upload Step

URL List (Text or CSV)

The simplest option: upload a text file with one URL per line, or a CSV file with a single url column.

Text File (.txt) – one URL per line:

https://example.com/
https://example.com/category/
https://example.com/page-1
https://example.com/page-2

CSV File (.csv) – single url column:

url
https://example.com/
https://example.com/category/
https://example.com/page-1
https://example.com/page-2

Requirements:

Use full absolute URLs (including https://).
One URL per line (for text) or one per row (for CSV).
Maximum file size and URL limits depend on your plan.

Sitemap XML

Standard XML sitemap format. You can export this from most CMSs or SEO tools.

<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
  <url>
    <loc>https://example.com/page1</loc>
    <lastmod>2024-01-15</lastmod>
    <changefreq>weekly</changefreq>
    <priority>0.8</priority>
  </url>
  <url>
    <loc>https://example.com/page2</loc>
  </url>
</urlset>

Tip: You can upload either a root sitemap or a sitemap index file. We’ll automatically follow nested sitemaps where supported.

Google Search Console Data (CSV)

Export from GSC with the following columns:

url,query,clicks,impressions,ctr,position,date
https://example.com/page1,"keyword 1",150,5000,0.03,12.5,2024-01-15
https://example.com/page1,"keyword 2",75,2500,0.03,18.2,2024-01-15
https://example.com/page2,"keyword 3",200,8000,0.025,8.3,2024-01-15

Tip: Export data with page + query dimensions for best results. Include at least 3 months of data for meaningful insights.

Page Embeddings (JSON or CSV)

JSON Format:

{
  "https://example.com/page1": [0.123, -0.456, 0.789, ...],
  "https://example.com/page2": [0.234, -0.567, 0.890, ...]
}

CSV Format:

url,embedding_0,embedding_1,embedding_2,...
https://example.com/page1,0.123,-0.456,0.789,...
https://example.com/page2,0.234,-0.567,0.890,...

Note: Embeddings can be generated using OpenAI's API, Sentence Transformers, or similar tools. Typical dimension: 384–1536.

Internal Links (CSV or JSONL)

Basic Format (CSV) – Required Features

Standard CSV file with internal linking structure:

source_url,target_url,anchor_text,follow
https://example.com/page1,https://example.com/page2,"Link text",dofollow
https://example.com/page1,https://example.com/page3,"Another link",dofollow
https://example.com/page2,https://example.com/page1,"Back link",dofollow

CSV Format provides: Basic link analysis, link metrics, cluster relationships.

Advanced Format (JSONL) – Enhanced Analysis

JSONL format (one JSON object per line) with anchor text embeddings for advanced analysis:

{
  "SourceUrl":"https://example.com/page1",
  "Outbound":[
    {
      "TargetUrl":"https://example.com/page2",
      "AnchorText":"Comprehensive Guide to Topic",
      "AnchorEmbeddingBytes":"YQwLPAf3SjwdzJG8...",
      "ContextEmbeddingBytes":"do0CPd4o8TwIGsA8...",
      "FullContextEmbeddingBytes":"CocCPSKE8TwD3b88...",
      "Context":"For more information, see our guide.",
      "HeadingContext":"Getting Started > Installation",
      "FullContext":"Section: Getting Started. For more...",
      "Position":1,
      "IsDoFollow":true,
      "ContextType":"p",
      "LinkPlacement":"editorial"
    }
  ],
  "Metadata":{
    "TotalOutbound":1,
    "LastUpdated":"2024-01-15T12:00:00Z",
    "EmbeddingModel":"text-embedding-3-small",
    "EmbeddingDimensions":1536
  }
}

JSONL Format provides:

Generic anchor detection ("click here", "read more", etc.)
Semantic mismatch detection (anchor/context vs target)
Context analysis (placement + surrounding text)
Quality scoring for each link

How to Generate JSONL Format

Tools that can extract this data: Screaming Frog, Sitebulb, custom crawlers, or contact us to use our built-in Page Links Service with your embedding provider.

Recommendation: Start with CSV for basic analysis. Upgrade to JSONL with embeddings when you need detailed anchor text quality and semantic mismatch detection.

Format Comparison

Feature	CSV Format	JSONL Format
Basic Link Metrics
Cluster Relationships
Generic Anchor Detection
Semantic Mismatch Analysis
Context Quality Scoring
Link Placement Analysis
Heading Context
File Size	Small	Larger
Setup Complexity	Easy	Moderate