CAPE: The Anonymous Network Built for Dog Trainers

CAPE: The Anonymous Network Built for Dog Trainers Who Are Sick of Facebook

Canine Advancement through Professional Exchange (CAPE) is an invitation-only platform where dog training and behaviour professionals can talk shop without their name, their business, or their reputation ever being on the line.

At a Glance

• CAPE is a private, invite-only platform for dog training and behaviour professionals — not the public.

• Every member posts under a pseudonym that changes with every conversation, so nobody can be tracked, targeted, or built into a "personality" across the platform.

• Conversations are organised as branches, trees, and a forest instead of a flat feed — nothing gets buried by an algorithm.

• Good answers are automatically saved to a searchable Resources library so knowledge doesn't get lost.

• An AI-powered tool called Answer Convergence detects when different professionals independently agree — and a whole "bias program" exists specifically to stop that agreement from being mistaken for consensus.

• Members can optionally add a professional profile (certifications, methodology, practice context) that makes the AI's bias-detection more accurate — never to rank, target, or expose individuals.

Why CAPE exists

CAPE started as a direct reaction to what dog training discussion looks like on Facebook: infighting, tribalism, self-promotion, marketing disguised as advice, and good conversations that vanish the moment the algorithm loses interest in them. Here's how the founder put it in the post that started it all:

"A thing that allows professionals to talk about their profession without being targeted by their name or their business, a space that is genuinely safe... I just want you to feel safe enough to have genuine conversations because some of my own conversations in this weird canine world push back against what is wrongly assumed to be fact."

That's the whole design brief. Every feature below — the anonymity, the lack of a public feed to perform on, even the AI tooling — exists to serve that one goal: make it safe enough for the professionals who usually stay quiet to actually say what they think.

In short: CAPE isn't trying to be a bigger or better social network — it's trying to be the opposite of one.

How conversations actually work: branches, trees, and the forest

Instead of a feed, CAPE organises discussion as a living structure:

• Every conversation is a branch. Replies can sprout their own branches partway through — not just at the end — so a side-discussion doesn't derail the original thread.

• A group of branches tracing back to one starting conversation is a tree.

• All the trees together make up the forest — the whole platform.

You don't decide in advance where something belongs. You reply to whatever prompted you, and the structure builds itself.

In short: CAPE replaces the endless scroll with a structure that actually preserves how conversations branch and evolve.

Pseudonyms: the mechanism that makes it safe

This is the direct answer to "talk about your profession without being targeted by your name or business."

Members sign up with a real name, but inside every conversation, they appear under a system-generated pseudonym — not one they choose. That pseudonym holds for the length of one conversation only; a new conversation means a fresh identity. Nobody can build a following, a target, or a reputation off watching how someone else posts.

In short: identity resets with every conversation, so ideas get judged on their own merit — not on who said them.

Private reactions, and the one deliberate exception

Instead of likes, every post carries a small set of private reaction "sensors" — a signal only the presser ever sees, for things like "this changed how I was thinking" or "I'm not sure I agree."

There's exactly one exception: the "great answer" reaction. Enough distinct people pressing it makes a small marker appear — "found helpful," "found helpful by several," or "widely found helpful." Two things worth being precise about:

• Who can see that marker exists: every member viewing that conversation.

• Who can see who pressed it, or the exact number: nobody — not other members, not moderators, not admins, ever.

In short: magnitude can be public; identity never is — not even to the platform's own admins.

Nothing good gets lost: fruiting, seeding, and the Resources library

When a conversation produces something genuinely useful — a resolved disagreement, a well-tested answer — it gets automatically "harvested" into a searchable Resources library, so good knowledge doesn't just scroll away and disappear the way it does on Facebook.

Occasionally, several separate conversations keep circling the same underlying idea. When that happens, CAPE can propose planting a brand-new tree to bring that convergence together — and unlike almost everything else on the platform, voting on these proposals is fully public: everyone can see the count and the threshold.

In short: valuable answers are catalogued automatically instead of vanishing into an algorithm's memory hole.

The AI features: Answer Convergence and Sensemaking

CAPE uses two admin-only AI tools to make sense of the forest at scale. Neither one ever changes what a member sees, and neither attaches a score, badge, or ranking to any person or post.

Answer Convergence looks across separate conversations for cases where different professionals, independently, land on similar advice. It uses AI embeddings (Voyage AI) to compare answers by meaning rather than exact wording, groups similar answers into clusters, and scores each cluster into a plain-language confidence band. Optionally — and off by default — a small, cost-gated AI model (Claude Haiku) can infer, in one sentence, what question a cluster is collectively answering.

Sensemaking is a separate curation tool that helps admins see the shape of the whole forest: which conversations cluster into topics, which have drifted off on their own, and which two conversations should probably be connected but aren't.

In short: the AI's job is comparison and clustering, never generation or judgment — and it never sees a real name.

Why agreement isn't the same as correctness — and why that matters

Here's the core caution baked into the entire system: independence between posters can never be fully verified. People who share a training background, a mentor, or a trending source can converge on the same answer for reasons that have nothing to do with that answer being right — the exact same failure mode that makes Facebook's "everyone agrees" so misleading in the first place.

That's why a hard rule runs through every AI feature on CAPE: nothing here ever recalculates, corrects, downgrades, or adjusts a convergence score. Every safeguard below is additional context next to the numbers — never a thumb on the scale.

In short: the goal isn't a "correct" score — it's an honest one.

Why CAPE asks for a professional profile (and why it's worth filling in)

This is worth being completely clear about: the profile isn't there to target or rank any organisation, methodology, or individual. CAPE never names an org or approach anywhere in its analysis, and it never will.

What it's actually for is making the convergence signal itself trustworthy. A raw count of "5 people said the same thing" has exactly the same weakness Facebook has, if all 5 happen to come from the same training lineage repeating what they were taught. That's not independent agreement — it just looks like it. Without knowing anything about who's in a cluster, the system genuinely can't tell the difference between "the field converged on this" and "one school of thought is well represented among today's posters."

An optional profile — certifying organisations, methodology, practice context — is what lets the system tell those two things apart. Filling it in doesn't touch anonymity inside conversations in any way, and admins only ever see aggregate counts (e.g. "3 distinct methodologies represented"), never which organisation or approach appears, and never below a minimum group size.

In short: every profile filled in makes the whole platform's bias-detection more accurate for every member — it's a small, one-time act that helps make the system harder to fool the way an algorithm fools people every day.

The full list of safeguards (13 bias types, three ways of handling them)

CAPE's design process identified 13 distinct ways "agreement" could be misleading and built a named mitigation for each:

Shown as context on every cluster (the things a profile directly helps with): training/methodology diversity, practice-context diversity, when the agreeing posts were made (trend bias), how many distinct people originally asked the underlying questions (anchoring bias), how concentrated a cluster's authors are elsewhere in the forest (clique/network bias), whether a question is more values-based than factual (ethical/normative bias), and a scrutiny flag for breed-level generalisations.

Disclosed permanently as "what this tool cannot see": the corpus is active posters, not the whole profession; confident writing clusters more easily than careful hedging; commercial affiliations are self-disclosed only and never shown; liability-safe advice converges for reasons that have nothing to do with what actually helps; shared mentor lineage is invisible by deliberate privacy choice.

Handled structurally, by how CAPE already works: rotating pseudonyms mean no one can build authority others simply defer to, and the "great answer" reaction can only ever add corroboration to a cluster, never subtract from it.

In short: every bias the design team could name got either a built fix, an honest disclosure, or a structural answer — nothing was left unaccounted for.

How to join CAPE

CAPE itself is invite-only — that's deliberate, for all the reasons above. If you'd like an invite, get in touch via the WhatsApp button on this website and ask; invites go out directly, not automatically.

In the meantime — yes, ironically — there's also a public CAPE Facebook group. It's the easiest place to follow updates, ask questions before you commit, or just see what this is about before requesting an invite to the platform itself.

In short: follow the Facebook group for updates, message via WhatsApp on this site for an invite to the platform.

Frequently asked questions

Is CAPE public, like Facebook groups?

No. CAPE is invitation-only, built specifically for dog training and behaviour professionals rather than the general public.

How do I join CAPE?

CAPE is invite-only, so you'll need to request an invite — the quickest way is via the WhatsApp button on this website. You can also follow the CAPE Facebook group for updates and a preview of what the platform is about before you request one.

Can other members see my real name?

No. Inside conversations, every member appears under a system-generated pseudonym that changes with every new conversation.

Does CAPE ever show who reacted to a post?

No, never — not to other members, moderators, or admins. The only public signal is a vague magnitude marker on the "great answer" reaction, and even that never reveals identity or an exact count.

Why does CAPE ask for a professional profile?

To make its AI bias-detection tool (Answer Convergence) more accurate — specifically, to tell genuine cross-organisation agreement apart from agreement that's really just one training lineage well represented among current posters. It's optional, never used to target or rank anyone, and admins only ever see aggregate counts, never which organisation or methodology is behind them.

Does the AI ever change a member's post or score their advice?

No. CAPE's AI tools only compare and cluster already-posted content to help admins understand patterns at scale — they never generate content on a member's behalf and never assign a score, rank, or badge to a person or post.

What happens to a good answer once it's posted?

Once enough distinct members mark it as helpful, it's automatically saved into a searchable Resources library so it stays findable instead of disappearing the way good advice does in a fast-moving social feed.

CAPE is built and maintained as an ongoing project. Features described here reflect the current build and may continue to evolve.