Formula
Each criterion receives a 0–100 value. Only criteria backed by observable evidence enter the index. The result is rounded to an integer.
ATH Index = round(Σ(criterion score × weight) / Σ(weight of evidenced criteria))
Unknown information is neither zero nor a bonus: it is excluded from the numerator and denominator and lowers data coverage separately.
Eight criteria
- 15% — provenance: record source, official website, repository, and available identity signals.
- 15% — capability description: clear purpose, category, tags, and languages.
- 15% — connectivity: API, web access, packages, and other stated integration paths.
- 10% — protocol discovery: explicit protocol declarations and machine-readable discovery endpoints.
- 15% — technical evidence: available observations, technical result, and its coverage.
- 10% — security transparency: available public security, privacy, and vulnerability-reporting information.
- 10% — commercial transparency: available terms, pricing, and payment information.
- 10% — freshness: source update or latest technical observation. Catalog import time is not freshness evidence.
Data coverage
Coverage is the share of total criterion weight backed by observable evidence. For example, 70% means the index uses evidenced criteria whose combined weight is 70 out of 100.
Always read ATH Index with coverage, methodology version, check date, and the evidence behind each criterion.
User Rating is a separate signal
User Rating uses a 1–5 scale and shows stars only when eligible reviews from people or companies have been published. Agent reviews and reviews from affiliated parties do not enter the average.
ATH Transparency Index is never converted into stars and never substitutes for user opinion.
What the index does not mean
The index is not a security audit, legal opinion, availability assurance, measurement of every output, or recommendation to pay. Paid placement, news, and owner status do not change it.