Editorial policy

ChocolateHQ is a structured reference. It holds records about chocolate types, cacao origins and varieties, production and preparation methods, flavour, pairings, history and health evidence, and it connects them with authored relationships rather than inferred ones.

It is not a recipe blog, not a brand review site, and not a manufacturer database. Those are three different products with three different failure modes, and being none of them is a decision rather than an omission.

This site does not review makers, rank bars or publish scores. There is no rating field anywhere in its data model.

The demand for “best chocolate brand” is real and this site declines it deliberately. Answering it would require either a rating system with no defensible basis or an affiliate-shaped list, and both would compromise everything else here. The comparison and tool sections answer the underlying need — which one suits this use — without naming a winner.

Nothing here is health advice or dietary advice, and there is no field in the data model that can hold a recommendation.

Every health topic states what the evidence shows and, in a section of equal weight directly beneath it, what the evidence does not show. That second section is required by the schema. Almost every serious failure in consumer health writing about chocolate is an overreach past a real finding rather than an invented one, and putting the boundary next to the finding is the only structural defence against it.

Where a quantity is inseparable from a claim, the amount is published with it. A finding about cocoa flavanols quoted without the intake it rests on is misleading in a predictable direction, because the amounts used in research are frequently not reachable by eating chocolate without also consuming a quantity of sugar and fat the same evidence says nothing favourable about.

A short list of phrases is prohibited outright in published prose and enforced automatically: cures, prevents cancer, detoxifies, boosts immunity, superfood, miracle, clinically proven and guaranteed to. None of them can be true of a food at any dose a reader will consume.

Facts about the world carry a confidence marker and their sources. Structural decisions — a route, a publication status — do not, because those are things this site decides rather than things it can be wrong about.

Where a widely circulated figure has no traceable original, this site records that it is unavailable rather than repeating the figure.

Two examples of that rule in practice: the share of world cacao production that is Criollo is quoted everywhere and appears nowhere here, and the cadmium limits in EU regulation are described as existing without their numbers being printed, because the regulation has not been re-read in this pass. See sources and evidence.

Not every page needs an image, and most pages here do not have one. Where a brief is unfulfilled the section is omitted entirely — no placeholder, no grey box, no stock photograph of unspecified brown squares.

An image is briefed only where it carries information the prose cannot. Several of the briefs in the catalogue are diagrams rather than photographs for that reason: the tempering curve and the production chain are things a reader genuinely cannot picture from a paragraph.

Synthetic imagery may never illustrate a named origin, variety, process step, defect or cultural practice. Any image of people at work in producing countries, of ceremonial cacao use, or from the colonial period requires an explicit cultural sensitivity note before it can be sourced.

Records are filed as documented history, living practice or contested narrative, and the last of those requires a written statement of what is disputed. Chocolate’s popular history is full of tidy dateable anecdotes that survive because they are better stories than the truth, and several of them have been repeated in chocolate advertising for over a century.

Practices with present-day practitioners are written in the present tense and are not described as survivals.

No advertising will ever appear on a health page, a myth page or a tool. Those page types are permanently ineligible in code, independently of any configuration.

No advertising appears anywhere on the site at present. When it does, the health surface and the tools will remain excluded, and the exclusion is enforced by a local gate that runs before any provider is consulted — so an outage or a misconfiguration cannot open it.

Records carry a last-reviewed date. If something here is wrong, the correction goes into the record rather than into a note appended to it, and the source register keeps the superseded material rather than overwriting it.