Agora is the front door to a family of coordination infrastructure.

The coordination failure Agora addresses does not stop at the meeting. A coalition that has documented exactly what its community needs still has to find an open bed tonight, still has to turn its evidence into an application a funder will read, and still has to carry one person through the same fragmented system it just finished mapping. The organizations doing that work still run on spreadsheets. Those are four more infrastructure problems, and they are the ones this company intends to build next.

Intends is the operative word, and this page is written so that it stays visible. One product is in the field. Of the four below, one is specified and being validated, one is designed with its contract to Agora already built, and two are commitments with nothing behind them yet except a reason. Agora’s status table applies the same discipline to what is finished.

Start a conversation The architecture they share

Four products, and exactly how far along each one is

Every card opens by saying how far along it is, because everything after that sentence describes something that mostly does not exist yet. Read in order, the claims get weaker. That is the honest shape of a roadmap, and it is why the last card is the shortest.

Poros

From the Greek for “way through.”

Poros is specified and currently being validated. None of it is built. It is a recovery capacity registry and warm-handoff network for the post-clinical recovery layer: recovery residences, faith-based residential programs, outpatient slots. This is the layer bed registries structurally cannot reach, because the providers in it are small, largely unlicensed, frequently faith-based, and answerable to no system that collects their availability. A coalition can know exactly what a person needs and still have no way to find out which door is open today.

What is specified is the registry, the handoff, and the freshness model. What is being validated is the only thing that decides whether it can work at all: whether providers in that layer will keep their own capacity current, for a network they are members of rather than customers of.

Design commitments already fixed

freshness-firstStale capacity data is suppressed rather than caveated. A registry that shows a three-week-old number with a timestamp beside it has not disclosed a limitation; it has moved the failure onto the person making the call at nine at night. If we cannot say a bed is there now, we do not say it.
flat membership fees, never per-placement feesMembership costs what membership costs, and it does not move when a placement happens. This is an anti-kickback posture stated as a design principle rather than as legal analysis. We are not offering a reading of what the law requires of anyone. We are saying that no part of the revenue model is allowed to pull against sending a person to the right door.

Convocant Grant Platform

The receiving end of an export Agora already writes.

The grant platform is designed, and one piece of it is already built. It converts a coalition’s published, evidence-backed briefings into structured grant-application material, closing the loop from community conversation to documented need to funded response to tracked outcome. The application stops being a writing exercise performed from memory in the two weeks before a deadline, and becomes a view onto a record the coalition has been building all along.

The contract between the two halves is the part that exists. An Agora briefing can be exported only after a named reviewer has published it, the export is versioned, and it is logged as a disclosure like any other. What is designed rather than built is everything on the far side of that contract.

Phoreo

From the Greek for “to carry.”

Phoreo is on the roadmap and nothing has been built. It would handle case-level navigation, carrying one person’s journey across the same multi-sector system the coalition coordinates, with case-level barrier-to-entry tagging on the shared Phragma taxonomy.

That shared taxonomy is the whole argument for building this here rather than as a separate thing. A barrier a navigator hits in one case and a barrier a coalition names in a planning session are the same barrier, and if they are tagged from one versioned vocabulary they can be counted together.

Oikos

From the Greek for “household.”

Oikos is on the roadmap and nothing has been built. It would be operations infrastructure for the nonprofits doing the work, meaning the organizations that carry the commitments a coalition makes, between the meetings where it makes them.

This card is short because Oikos is the least specified thing on this page, and padding it out would only make it sound further along than it is.

On the status table on the Agora page, all four of these products sit in a single roadmap row. The statuses above say where inside that row each one is: what has been specified, what has been designed, and the one piece that is genuinely built, which is the export contract on the Agora side rather than the Grant Platform itself. No product on this page carries a launch date, and the section below says why not.

Illustrations: abstract motifs, one per product. No participant data.

One posture across the portfolio

A portfolio is only worth the name if the second product costs less than the first because of the first. These six commitments are where that saving comes from, and they are why a product on this page is a roadmap item rather than a rewrite. None of them is a feature any of these products has to go and build.

consent-record-first
Consent is a record with its own lifecycle, written before anything is captured. It is not a checkbox recalled afterwards and not a paragraph in a terms page. Every product that touches a person’s information starts from that record, and the capture path cannot begin without one.
per-record keys
Encryption happens in the application rather than only at the disk, with a separate data key for every record. A database taken whole yields ciphertext, whether it was stolen, subpoenaed or simply misplaced, and destroying a record’s keys makes it permanently unreadable by anyone, including us.
append-only audit log
Every sensitive read, disclosure and consent change is written to a hash-chained, append-only log holding metadata rather than content. Verification re-derives the chain, so tampering is not merely against the rules. It is arithmetically visible to anyone holding the log.
provider-swappable AI
Model providers sit behind an interface, and no product is written against one vendor’s API. Replacing a provider is a configuration change rather than a rewrite, and a provider with no recorded agreement covering protected data cannot be sent any.
human-gated publication
Nothing a model generates reaches a public surface, a funder or a filing without a named person releasing it. There is no path through any of these products that publishes on its own, not behind a flag and not for a trusted tenant, and there is not going to be one.
versioned taxonomies
The shared vocabularies, the Phragma barrier taxonomy first among them, are versioned configuration rather than code. A finding tagged in Agora means the same thing when Phoreo reads it, and a taxonomy revision is an event with a version number rather than a silent change in what last year’s data meant.

Every product inherits this spine. That is the reason there is a portfolio at all.

Each of these is a claim about the Agora codebase as it runs today rather than an aspiration for the portfolio. Every one is stated as its mechanism, with the identifiers a technical reviewer would grep for, on the trust page.

We build in order, and each step is gated

This follows the same discipline as the product milestones. A thing is not started because it is on a list. It is started because the layer under it is carrying weight and somebody has demonstrated they need the next one. What follows is an order and its gates, which is the only part of a roadmap we can honestly publish.

Now Agora, in the field with real coalitions. Everything else is gated on it being load-bearing for them first.
Next Poros, gated on validation. That means providers in the post-clinical layer keeping their own capacity current, at a rate that makes a freshness-first registry usable rather than empty.
After that The Grant Platform, gated on coalitions having enough longitudinal record for an export to be worth more than a blank application form.
Later Phoreo, then Oikos, gated on the layers beneath them carrying weight and on somebody needing them badly enough to say so.

Nothing on this page carries a launch date, and that is deliberate rather than evasive. A date is the easiest thing here to write and the only thing on the page we could not stand behind. Every one of these gates is a fact about other people’s behavior, and we do not get to schedule those.

The order is not ours to fix alone.

Every gate above is waiting on evidence, and a community telling us which of these it needs first counts as evidence. If the answer for you is the capacity registry rather than the meeting platform, we would rather hear that now than work it out from the roadmap later.

Start a conversation See what is built today