# Open Source Stats
Searchable comparison tables of verified statistics. Every figure sourced with provenance and collection method.

> Base URL: https://opensourcestats.com

## AI Coding Tools — GitHub Adoption (ai-coding-tools-github)
- Latest comparison: https://opensourcestats.com/compare/ai-coding-tools-github/2026-10/
- Machine exports:
  - JSON: https://opensourcestats.com/data/ai-coding-tools-github/2026-10.json
  - CSV: https://opensourcestats.com/data/ai-coding-tools-github/2026-10.csv
- Datapoint groups:
  - GitHub Stars: Total stargazers count (e.g. /s/anthropics-claude-code-github-stars/)
  - GitHub Forks: Total repository forks
  - Open Issues: Open issues count
  - Stars Per Day: Star velocity since creation
  - Fork-to-Star Ratio: Forks divided by stars percentage

## History, growth and rank changes — the figures no single API call returns
- Growth, rank change and overtakes: https://opensourcestats.com/history/
  Star and fork growth over 30/90/180/365 days, rank movements and crossover dates, computed from
  dated Wayback Machine captures of each repository page (2,268 captured points across three
  counters). Not interpolated: each figure names
  the capture date it used. This is reconstructed data, kept separate from measured readings.

## Dated total across the tracked set
- Combined GitHub stars, all tracked tools: https://opensourcestats.com/daci/
  One dated sum of stars across every tracked repository. Deliberately described rather than
  branded: 'AI Coding Index' is already used by another publisher, and a name that collides
  with an established index is worse than no name.

## Downloadable dataset
- Dataset documentation, field dictionary and JSON Schema: https://opensourcestats.com/datasets/reconstructed-history/
- Points as CSV: https://opensourcestats.com/data/reconstructed-history.csv
- Points as JSON, self-describing: https://opensourcestats.com/data/reconstructed-history.json
- JSON Schema (2020-12): https://opensourcestats.com/data/reconstructed-history.schema.json
  Reconstructed from dated Wayback Machine captures - sparse, never interpolated, and NOT
  measured readings. Validate against the schema rather than parsing hopefully.

## Measurement and crawl activity
- Crawl signals (request counts, not index entries): https://opensourcestats.com/signals/
- Target-query rankings (what is measured, what is blocked): https://opensourcestats.com/ranking/

## Method, definitions and limitations
- Methodology: https://opensourcestats.com/methodology/
  States the exact endpoint and fields read, how each figure is derived, which repositories
  failed to resolve, and what these numbers cannot tell you.
- GitHub Stars: https://opensourcestats.com/metrics/github-stars/
- GitHub Forks: https://opensourcestats.com/metrics/github-forks/
- Open Issues: https://opensourcestats.com/metrics/open-issues/
- Stars Per Day: https://opensourcestats.com/metrics/stars-per-day/
- Fork-to-Star Ratio: https://opensourcestats.com/metrics/fork-ratio/
  Definition, source field, formula (for derived figures) and known pitfalls per metric.

## Full Text & Discoverability
- Full text index: https://opensourcestats.com/llms-full.txt
- Data index: https://opensourcestats.com/data/index.json
- Sitemap: https://opensourcestats.com/sitemap.xml