README.md

---
title: CC Backlink Explorer
emoji: 🔗
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 5.50.0
app_file: app.py
pinned: false
license: mit
---

# CC Backlink Explorer

A Majestic/Ahrefs-style **backlink & rank-booster** tool built entirely on the
[Common Crawl](https://commoncrawl.org) domain-level web graph
(release CC-MAIN-2026 Apr–May–Jun: **121M domains, 3.9B links**).

## Features
- **Backlink report** — referring domains, harmonic centrality rank, PageRank, and the
  top referring domains ranked by their own authority.
- **Link gap / rank booster** — high-authority domains that link to your competitors but
  not to you: the highest-leverage targets to raise your harmonic/PageRank.

## Data
- `domain_vertices.parquet` and `domain_ranks.parquet` are downloaded locally at startup.
- `domain_edges.parquet` (3.9B edges, sorted by `to_id`) is queried remotely from the HF
  Hub via DuckDB `httpfs` with HTTP range requests, so each backlink lookup fetches only
  the relevant row groups.

Note: the Common Crawl web graph is host- and domain-level. Page/URL-level backlinks are
not available here (that requires processing WAT files).