Expertise is not scarce. Access to it is.

There are people in every city and every vertical who have already solved the problem you are currently stuck on, and the distance between you and thirty minutes of their time is almost never about money. It is about not knowing who they are, having no legitimate reason to contact them, and — from their side — no sane way to field requests from strangers without it consuming the working week.

CompassVia is the platform built to close that gap: a gated expert-advisory community with a curated directory of local and vertical experts, and a booking layer that makes their time available in formats that actually scale.

This is the largest and most interconnected build in this set, and it is worth walking through what "platform" means in practice, because the word gets used loosely.

Three ways to get access, not one

The obvious version of this product is 1:1 consulting bookings. CompassVia does support those, but leading with them would be a mistake, and the platform is structured to avoid it.

An expert's calendar is the hardest-capped resource in the system. If the only format is one-to-one, supply runs out immediately, prices rise until access is rationed by budget, and the community half of the proposition dies. So access comes in three shapes:

Group office hours — one expert, many members, the same hour. Most questions people are afraid to ask are questions several people in the room have.
AMAs — broader still, and asynchronous-friendly, so an expert's time produces an artifact rather than evaporating when the call ends.
1:1 sessions — for the situations that genuinely require privacy and specificity.

Each of those is a different scheduling problem, a different capacity model, a different notification pattern, and a different pricing question. Supporting all three is what makes the economics work, and it is also a large part of why the build is platform-scale rather than app-scale.

The gate is the product

CompassVia is gated. Tiered paid membership controls access, and that is not a monetisation layer bolted on top of an open directory — it is load-bearing.

An ungated expert directory becomes a lead list. Experts get cold-pitched, the good ones leave, and the directory decays into a listing of people who no longer participate. The gate is what makes it safe for a genuinely good expert to be listed at all, because everyone on the other side of it has a membership, a profile, and something to lose.

That means the membership tier is not a pricing table. It is an authorization axis that has to be consulted almost everywhere: who can see which experts, who can book which formats, how many sessions a tier includes, what happens at the boundary. Tier logic that lives in the UI instead of the API is how gated platforms leak — and this one runs a separate API service at api.compassvia.com, which is precisely the boundary where that enforcement belongs.

Two-sided means two products

Members are only half the system. Experts have their own path through it: an application to join, a public profile, availability to manage, sessions to run, and money to be paid.

Every marketplace discovers the same thing — the supply side is a full product in its own right, and it is usually the one that gets under-built. Experts here get applications and public profiles, and critically, revenue share via Stripe. Payment is not an afterthought. If an expert cannot see what they earned and when it arrives, they disengage, and a platform without engaged experts is a directory.

Stripe subscriptions for members plus revenue share to experts means money flows in on one schedule and out on another, against sessions that may be group or individual, attended or cancelled. That reconciliation is one of those areas where the requirements sound simple in a meeting and turn out to have a dozen states.

Video that works on the first try

Sessions run on Daily.co, with pre-call quality testing before a member joins.

That pre-call check is a small feature that reveals how seriously the product takes its own failure modes. The single worst experience this platform can deliver is a paying member joining a booked session with a busy expert and spending the first six minutes discovering their microphone does not work. It burns the member's money, the expert's time, and the platform's credibility simultaneously — and it cannot be recovered afterwards.

Testing the connection before the door opens moves that failure to a moment where it costs nothing. Building on an established video infrastructure rather than raw WebRTC is the other half of that decision: video quality is a problem with no ceiling and no advantage in solving yourself.

Security proportional to the stakes

The platform carries TOTP two-factor authentication and SMS verification, alongside an invitation system.

For a consumer content site that would be overkill. Here it is proportionate. Accounts hold paid memberships and booking rights; expert accounts are tied to earnings. Verification also does community work — it raises the cost of creating a fake identity, which is the main way a curated professional network gets degraded.

Invitations serve the same end from the growth side. Growth through existing members preserves the curation that makes the directory worth paying for, rather than eroding it.

The parts nobody sees

Three more pieces round out the build and are worth naming because they are where platform projects tend to be underestimated.

A playbooks and templates library captures knowledge that does not need a live session at all — the highest-leverage content in the system, because it answers questions without consuming anyone's calendar.

Timezone-aware reminders sound trivial and are not. Every booking involves at least two people, possibly a room full, across arbitrary timezones, with daylight-saving transitions that move independently by country. Getting this wrong means someone misses a session they paid for.

Admin tooling is the difference between a platform that can be operated and one that requires a developer for every expert application, refund, and support question.

Why this one is a platform

CompassVia is a client-rendered application backed by a separate API — a Vite single-page app talking to a service at its own subdomain. That split is the right architecture for something this stateful, and it is a fair marker of where a project crosses from website-plus-app into platform: when the front end is genuinely an application, and the business logic needs a home that is not the browser.

Multi-party permissions, two-sided economics, real-time video, money moving in two directions, and identity verification — any one of those is an app. Together they are a platform, and pricing a build like this as anything else does nobody any favours.