Every reading, as structured files
With each final reading, everything needed to check a score is published: the ranking, the cell-level record of who was named where, the prompt set and the universe list. Free to download and use, with attribution, for checking, reporting, teaching and research. Answer text is withheld by design.
Files per reading
Files are published with each final reading, after the human check: the same five files for every market × reading, at the paths below. No reading has final files yet. The provisional home-builders reading (November 2026) is shown on its ranking page; its files publish when the reading is signed off.
Files are versioned; corrections publish a new version and keep the old one, with a dated note on the reading.
| File | Contents | Path |
|---|---|---|
| ranking.csv | One row per eligible business: rank, share, spread, prompts named in, engines naming, first positions, sources cited, change, relationship, status | /data/{market}/{YYYY-MM}/ranking.csv |
| cells.csv | One row per prompt × engine × run × business named: prompt id, engine, run, timestamp, position, source URLs. No answer text. | /data/{market}/{YYYY-MM}/cells.csv |
| reading.json | Reading metadata, headline, findings, method version, universe source and date, exclusions, analyst | /data/{market}/{YYYY-MM}/reading.json |
| prompts.json | The 17 prompts with pillar, funnel stage and rank-score flag, as asked | /data/{market}/{YYYY-MM}/prompts.json |
| universe.csv | Every eligible business with registry id, segment and resolution notes, as of the export date | /data/{market}/{YYYY-MM}/universe.csv |
Machine-readable entry points
- /sitemap.xml — every ranking, business, prompt and reading page, with the date each page was last updated.
- JSON-LD: Organization on business pages and DataCatalog on this page today. Dataset markup on reading pages, with links to the files above, is added when the first final reading publishes.
- /llms.txt — a plain-text summary of what the index is and where the data lives, written for AI crawlers. Published with the site's own domain; not yet live.
cells.csv columns
A business absent from cells.csv for a reading was not named in any cell. The published scores are reproducible from cells.csv and the position weights in the methodology; a mismatch is a correction request.
Licence
Data files are not licensed under CC BY. They are free to download and use for checking, reporting, teaching and non-commercial research, with credit: “Trust Index (LittPro Inc.), [market], [reading]” and a link to the reading. Redistributing the files or using them commercially needs a data licence from LittPro Inc. (see Terms). Readings, rankings and charts as published on these pages are licensed CC BY 4.0.
Until a reading is final, its figures are shown on the ranking page and labelled provisional; no file is published for a provisional reading.