About PawLabs

What PawLabs is

PawLabs exists because choosing a breed is usually done badly. People pick on looks, on a dog they met once, or on a list that ranks every breed as "loyal and loving" — and then discover in month three that the breed needs two hours of exercise they do not have.

So we do the boring thing properly. Every breed gets the same treatment: a full profile, the same 14 traits scored 1–10, honest weaknesses alongside the strengths, and a clear statement of who should not get it. Because the scoring is identical across breeds, you can put any two side by side and the numbers actually mean something relative to each other.

Today the site holds:

• 201 breed profiles, each with temperament, health, grooming, exercise and living-condition detail.
• 3 head-to-head comparisons for the match-ups people agonise over.
• A Breed Match quiz that scores your answers against every published breed rather than handing you the same five popular dogs everyone else gets.

We are independent. No breeder, kennel club, pet food company or rescue organisation pays for placement here, and none of them get a say in what a profile says.

Who writes this

Every breed profile and comparison on this site carries that byline. It is a team, not a pseudonym for one person, and it is not a stand-in for a veterinarian — we are researchers and editors, not clinicians. Where a claim is genuinely medical, we point you at your vet rather than pretending otherwise.

If you are a veterinarian, breeder, trainer or behaviourist and you want to review a profile, or argue with one, we would like to hear from you: contact@pawlabs.org.

How we research a breed

Each profile is built from the same source hierarchy, in this order:

1. Published kennel club breed standards. The recognised standards define what a breed is meant to be — size, coat, structure, original working purpose, described temperament. This is where the structural facts come from.
2. Veterinary and breed-health reference sources. Lifespan ranges and the conditions a breed is predisposed to are drawn from veterinary literature and breed health surveys, not from anecdote.
3. The breed's working history. What a breed was bred to do explains most of what it is like to live with. A dog bred to run all day is not going to be satisfied by a walk around the block, whatever a listicle says.

The 14 trait scores are assigned against that source material and deliberately spread across the range. If everything scored 8 out of 10 the scores would be worthless, so a breed that is genuinely hard to train gets a low trainability score and we say why.

How we use AI

We use AI, and we would rather tell you than have you work it out. Here is exactly where:

• Drafting profiles. Breed profiles and comparisons are drafted by a large language model working from the source material described above. Every draft is saved unpublished and goes through human review before it can appear on the site.
• Quiz matching. Your quiz answers are scored against every published breed by a deterministic scoring function first. AI then picks from that shortlist and writes the explanation. If the AI service is unavailable, the scoring alone produces your result and we label it as such on the page.

What AI does not do here: it does not decide trait scores without review, it does not publish anything on its own, and it is never used to invent a credential or a qualification. We do not claim veterinary standing we do not have.

One thing we are actively cleaning up: some older profiles are written in a first-person voice that reads like lived experience with a particular dog. That was a drafting style we have since dropped, because a research team should not write as though it has personally owned every breed. Newer profiles describe what breed data and owner reports show, in the second person. If you spot an older one that still reads the wrong way, tell us and we will rewrite it.

Editorial standards

• No pay for placement. No one can buy a higher trait score, a better quiz match, or a place in a comparison.
• Commercial relationships never touch the ratings. Trait scores come from breed standards and health data. Affiliate arrangements are described in full on our Affiliate Disclosure page.
• Weaknesses get equal billing. Every profile names who should avoid the breed. A profile that cannot say anything negative about a breed has not been researched.
• Health claims stay conservative. We describe predispositions and typical lifespans. We do not diagnose, recommend treatment, or suggest you skip a vet visit.
• Profiles get revised. When breed health data or a standard changes, or when a reader shows us we are wrong, the page changes and the modified date moves with it.

Corrections

We will get things wrong. When that happens we want to know, and we would rather fix it quickly than defend it. Email contact@pawlabs.org with the page, the specific claim, and — if you have one — a source. Substantive factual corrections are made to the page and the last-updated date is moved.

Full detail on what to include is on the Contact page.

How we make money

Running a site of this size costs money, and we would rather be straight about where it comes from than pretend it appears from nowhere. PawLabs is supported by affiliate commissions and advertising. Neither buys influence over the ratings or the quiz.

The mechanics — what an affiliate link is, how ours are labelled, and what it means for what you read — are set out on the Affiliate Disclosure page. What data we collect and who processes it is on the Privacy Policy page.

What this is not

Get in touch

Questions, corrections, press enquiries and data requests all go to contact@pawlabs.org, or use the routes on the Contact page. A real person reads it.