# The Observatory

> The independent reference on how AI and search surfaces name, cite and frame brands across premium markets. Read each observation within its stated scope.

Scope note: observations are readings of a named AI or search surface within a stated market, window and limitation. They are not endorsements, business-market rankings or market-share claims. Recorded model statements are distinct from checked facts and from measured metrics. Measured by AAI Kernel, published by The Observatory; The Observatory does not sell placement.

## Find a market or a reading

- [Markets](https://theobservatory.io/markets): choose a market type, then a destination; each room reads its market on its own terms.
- [All cards](https://theobservatory.io/observations): the dated archive of every published card, newest first; follow the pagination for older observations.
- [Rankings](https://theobservatory.io/rankings): measured ranking tables, each scoped to a surface and an observed window.
- [Spotlight](https://theobservatory.io/spotlight): one query, one surface, one answer moment.
- [DNC Lens](https://theobservatory.io/dnc-lens): a named model's recorded reading of a market, left uncorrected.
- [Model Watch](https://theobservatory.io/model-watch): comparisons between explicitly named model/surface readings.
- [Actors](https://theobservatory.io/actors): observed identity navigation; the source cards establish the evidence scope.

## Interpret a reading

- [Method](https://theobservatory.io/method): publication and measurement boundaries; how readings are scoped (#standards), the source protocol (#source-protocol), native AI perception (#dnc-note).
- [Query panel](https://theobservatory.io/method/panel): query-panel construction and run conditions (#run-metadata).
- [Challenges](https://theobservatory.io/method/challenges): limitations of the approach, in the Observatory's own words.
- [Reference](https://theobservatory.io/reference): concept explanations for reading the answer layer.
- [Glossary](https://theobservatory.io/reference/glossary): short canonical definitions; the stable term identity is /reference/glossary#<term>.

## Use or cite an observation

- [Press & citation](https://theobservatory.io/press): attribution and citation guidance.
- [Public intelligence library](https://theobservatory.io/press/citations): every public work — governed readings, Page Context, method, reference, actor records — searchable by type, market and date, each with RIS / BibTeX / CSL-JSON / JSON; the source page remains authoritative.
- [Feeds](https://theobservatory.io/press#feeds): RSS 2.0 and JSON Feed 1.1 per work type, market type and market, plus material corrections; excerpts and canonical links only.
- [Terms](https://theobservatory.io/terms): the public reuse scope and its exclusions.
Use the card's `#use-this-observation` section. Where a card offers them, `citation.txt` carries the complete public citation record and `data.csv` carries its published metric table (a published slice, not the underlying capture). Reuse: Free reuse, including commercial republication, under The Observatory’s Terms. Retain attribution, source link, scope and date.
Attribution that travels with a reused reading: "Published by The Observatory · Based on an AAI Kernel capture · governed by The Observatory editorial system" Usage policy: "Observed by The Observatory. Not an endorsement, recommendation, market-share claim or commercial ranking."

## Read dates correctly

An observed period, a publication date and a last-checked date mean different things. The observed window is when the surface was read; a publication date, where one is stated, is when the frozen card was published; "last checked" on a model reading is when its factual claims were checked. Permanent dated cards describe their recorded scope — do not treat a snapshot as the current state of a market; use the archive or the market room to find later readings. A card's structured data carries the observed window as the dataset's temporal coverage and states a publication only when one is governed; measured cards may expose a Dataset, evergreen pages do not.

## Apply the reference

- [Decision guides](https://theobservatory.io/decision-guides): interpretation frameworks for reading the answer layer.
- [Use cases](https://theobservatory.io/use-cases): applications of the public reading standard.

## Corrections and contact

- [Corrections](https://theobservatory.io/method#corrections): the correction and right-of-reply procedure.
- [Right of reply](https://theobservatory.io/right-of-reply): submit a relevant response or a correction request.
- [About](https://theobservatory.io/about): publisher context.
- [How The Observatory tracks brands in AI search](https://theobservatory.io/about/brands-in-ai-search): what is tracked, how often, and what comparison can support.
- [Contact](https://theobservatory.io/contact): contact routes.