A Gord Holdings companyRoadmapFlashyOS ↗
4.venturesA Gord Holdings company

The record

Ventures

Every company this studio ideated or inherited, each with the position it was started on — what was believed, what had to be true, and the single observation that would have shown the belief wrong. 33 theses, 21 of them originated here.

Each thesis carries a grade and the register the grade was read from. Several read Argued, which means the reasoning is sound, the thing exists, and nobody outside this estate has used it. That is published rather than hidden: a portfolio shown only at its best is one a reader is right to discount.
§ 01

The ladder

Four grades, each defined by a fact somebody outside this studio could check. A company moves up when a number moves, never when somebody decides it feels ready.

GradeWhat it meansCount
M4ProvenRunning against a population outside this estate, with a number that moves without our help.2
M3EvidencedRunning in production. Every consumer is inside the estate, which is the ceiling until somebody outside adopts it.9
M2ArguedThe reasoning is sound and falsifiable and the thing exists. Nobody has used it yet.20
M1Pre-thesisA real observation about the world. Not yet a company, and named here rather than dressed as one.2
§ 02

M4 · Proven

Running against a population outside this estate, with a number that moves without our help.

Metaverse Group2021 — realisedDated 2021 · to the year
InheritedRealised
Virtual land would become an asset class because worlds would develop economies inside them.

Basis · Realised. The falsifier occurred: virtual land repriced sharply in 2022 and most has not recovered, because there was nothing to do on it. The thesis was tested and answered — which is what `proven` means here, not that it was right.

BitMap Holdings2023 — the correction to Metaverse Group, by the same peopleDated 2025 · to the year
InheritedRealised
The defect in every platform metaverse was never the world; it was that the property inside it existed at the pleasure of an operator. Rebuild the position on a substrate where nobody can revoke, dilute or mint against it, and the scarcity becomes worth something a buyer can underwrite.

Basis · Realised: acquired by Skrybit in 2025 in a GDA-led transaction, with the record published at gda.group/transactions/skrybit-acquires-bitmap-holdings/. The falsifier half-occurred, and `enforceable scarcity` entered the canon as W4-003 because of it.

§ 03

M3 · Evidenced

Running in production. Every consumer is inside the estate, which is the ceiling until somebody outside adopts it.

Flashy Gold2022 — after the rails thesis met a distribution problemDated 2026-08-11
Inherited
A reward is only worth something if it can be redeemed for something outside the system that issued it, and the company that owns the redemption path owns the economics — not the one that owns the points.

Basis · claimyour.gold runs in production with a live redemption path and a per-hunter cap enforced in the format, not in a settings panel. Every hunter is reached through estate distribution; no external property has adopted the ledger.

Flashy Network2023 — when the ledger needed somewhere to settleDated 2026-08-14
Inherited
The group needed settlement it controlled, and the way to get it was not to compete with layer ones on consensus but to sit above them and be the only place where two chains reconcile.

Basis · Live at flashynetwork.com, settling for estate properties, with verification that runs in the reader’s own browser. The countersignature ratio it publishes reads 0 of 28: no counterparty outside this estate has signed a claim it makes.

Flashy Group2022–2024 — assembled over six transactionsDated 2021 · to the year
InheritedOperating
Three of the four conditions of a working digital economy could be held at once by acquisition, in an order where each asset makes the next one cheaper — and that the order is the whole argument.

Basis · Nine properties on one ledger, agent-operated, with the six transactions published tombstone by tombstone at gda.group/transactions/. Evidenced rather than proven because every user reached is still reached through estate distribution: adoption-signal.json reads independentAdopters 0.

FlashyOS2024 — when nine properties needed one another’s agentsDated 2026-08-26
Inherited
When agents act for organisations rather than for users, the scarce thing is not intelligence but accountable coordination: a way for two organisations to agree work, do it, and hold a record neither can quietly revise.

Basis · Live at flashyos.com and app.flashyos.com, with the estate as its own first tenant. adoption-signal.json records independentAdopters: 0 — the figure the whole roadmap turns on.

Countersign2026 — when the record needed proof a stranger could check without trusting either partyDated 2026-09-22
Originated here
A federated record is only worth building if a claim about another organisation stays attributed to whoever made it until THAT organisation signs — never asserted as settled by whoever is doing the claiming. A shared database makes the consent question disappear by construction, because whoever holds write access settles it; a blockchain makes checking expensive and public in the wrong direction, because everyone must be a party to verify anything. Countersignature keeps the claim, the signature and the check as three separable things a stranger can hold without trusting us, the asserter, or a validator set.

Basis · @flashyos/countersign ships a CLI (`init`, `validate`, `grade`, `audit`, `invite`, `countersign`, `verify`, `delegate`, `conformance`) with 29 test files and a 23-case conformance suite pinned by test. flashynetwork.com/countersignatures publishes the countersigned share live — 0 of 30 cross-boundary claims countersigned, 4 of 30 corroborated but not signed, read directly from the page on 2026-09-23. flashynetwork.com/notary is RFC 6962: inclusion and consistency proofs over 7,681 leaves from a dozen sources, hashes only, never the record. The backlog/1-to-intent/1 fold (flashy-network commit 4964045, 2026-09-22) reads six properties’ promoted backlog items — flashyos, flashygold, flashynetwork, claimyourgold, academy, flashyid — through `backlogToFragment`; the served `notary-log.json` was still generated 2026-09-22T20:33:44Z as of this writing, forty minutes before that commit landed, so the fold is shipped and tested but its leaves had not yet reached a served snapshot when checked live. Evidenced rather than proven: every consumer of the countersignature mechanism — asserter and verifier both — is still inside this estate.

Flashy ID2024 — when agents needed to prove who authorised themDated 2026-08-18
InheritedBuild
Identity failed least dramatically and matters most, and agents are what finally make it urgent: a workforce of software must prove what it is and who authorised it, to counterparties with no relationship to it.

Basis · Live at flashyid.com with the SDK published and the grant kernel enforcing attenuation. Every relying party is an estate property.

Flashy Academy2024 — when the ledger could pay people and not rank themDated 2026-07-09
InheritedBuild
A network whose asset is who-knows-whom needs a way for standing to be earned rather than bought, and teaching is the cheapest honest way to earn it.

Basis · Live at flashy.academy with the economy running. Every learner arrives through estate distribution.

ClaimYour.Gold2025 — when the ledger needed a population, not another integrationDated 2026-05-01
InheritedOperating
The redemption path is the product. A rewards system is believed the first time somebody receives something real, and never before.

Basis · In production at claimyour.gold — hunters, a partner tier and a live redemption path, promoted from main to production on a gated workflow rather than on somebody remembering. The strongest evidence in the estate that the ledger thesis holds, and still entirely inside estate distribution.

NFT BAZL2021 — the moment the category had money and no shared definitionDated 2021 · to the year
InheritedOperating
Convening an industry in public is the cheapest way to find out whether a category exists, before committing capital to it.

Basis · Operating, with the convening record held at gda.group. Relationships formed there persist across the estate; adoption-signal.json still reads independentAdopters 0, so none of them has yet adopted anything the estate publishes.

§ 04

M2 · Argued

The reasoning is sound and falsifiable and the thing exists. Nobody has used it yet.

The wallet authorization plane2026 — when FlashyOS agents needed real spending authority, not just an approval queueDated 2026-09-22
Originated here
An agent that can be given real spending authority is worth more than one that can only draft a request for a human to execute — but only if the organisation can still say afterward, to itself and to anyone who asks, exactly what was authorized, what happened, and why the two matched. Separating who-may-spend from what-just-happened is what makes that answerable; a faster approval workflow bolted onto a hot wallet does not.

Basis · @flashyos/signer, @flashyos/wallet-wdk and @flashyos/wdk published to npm at 0.1.0 on 2026-09-22, confirmed against the registry rather than the repository, with 91, 117 and 34 of their own tests passing per PR #137’s own table. `TESTNETS` in the signer is the whole list of chains it will construct a backend for — Base, Ethereum, Arbitrum and OP Sepolia, Nile and Shasta — and a mainnet chain id throws `MainnetNotEnabled` by construction, not behind a flag. In production, `authorizationKeys()` throws if `WALLET_AUTHZ_PRIVATE_KEY` is unset rather than minting an ephemeral key, so the plane cannot sign anything in production until an operator sets it — read from `packages/api/src/lib/wallet/authorization.ts`, not asserted from a doc. flashyos.com/wallet 404s as of 2026-09-23; PR #137’s own body says the public surface is Phase 34, gated on this merge and not yet shipped. Argued rather than evidenced: the thing exists, is tested, and runs on testnets — nobody, inside or outside the estate, has yet spent through it in production.

Flashy Mind2025 — when coordination needed memoryDated 2026-08-26
InheritedDesign
An organisation’s agents are only as useful as what the organisation can remember, and enterprise memory is a governance problem before it is a retrieval problem.

Basis · Status `design` in the portfolio register. A route on flashyos.com and no independent surface. The most crowded claim in the stack and the one with the least evidence behind it.

IntentMesh2026 — extracted from FlashyOS when the future tense turned out to be a formatDated 2026-08-31
Originated here
Organisations publish what they have done and never what they intend to do, so agents can only react. A public, decaying record of intention is the missing half of coordination.

Basis · Rung `section` on the flashyos incubation ladder. The gate to `property` is one organisation outside this estate running the format of its own accord; adoption-signal.json reads 0. 58 items filed across thirty repositories and 0 public.

AssetMesh2026 — the denomination layer, split out from the intent workDated 2026-09-02
Originated hereBuild
Real-world assets fail at denomination rather than at custody: the hard part is agreeing what one unit of a thing IS, in a way a stranger can check without trusting the issuer.

Basis · Rung `section` on the incubation ladder. rwa/1 is specified and the validator runs; zero issuers. The founding-issuer offer deliberately publishes no percentage, valuation or dollar figure, and a test greps the copy to keep it that way.

Magician2026 — the reach problem, once coordination workedDated 2026-09-01
Originated here
Coordination between organisations that already know each other is a solved problem. The unsolved one is reach: a federated way to be introduced to somebody you have no path to, without either party surrendering their graph.

Basis · magician.network is built and its four packages are on no registry. estate-packages.mjs reads them absent; @magician/concierge ships the activation checker this estate runs as its tenth standard, so the standard’s own checker cannot currently be installed.

RitualOS2026 — the present tenseDated 2026-09-02
Originated hereOperating
Records cover what shipped and what is intended, and miss what is practised. Recurring witnessed observance is a third tense, and unwitnessed practice should move nothing.

Basis · ritual/1 is specified, practised by org/flashyos, and measured by estate-ritual.mjs at 2 of 37 adoption with **zero** witnesses estate-wide. The copy-detection guard was written before the first witness exists, deliberately.

BuildOS2025 — when the coordination layer had no vertical to prove itself inDated 2026-08-23
Originated hereBuild
Construction and property development run on coordination between parties who do not share a system, which is the same shape as the agent problem one industry over.

Basis · Status `building`. buildos.com is registered at NameBright with a zone that answers nothing, so the property has never been reachable. The only holding whose thesis has never had a live surface.

NFT Capital2026 — after enough marketplaces had closed to answer the question with evidenceDated 2026-09-01
Originated here
Digital property needs an assay office before it needs a market: what holding a token actually confers, and what happens when the marketplace that issued it closes.

Basis · Built: 22 pages, charter, handshake, front door, card and icons. nft.capital resolves to a registrar parking page, so none of it has ever been served.

h3alth Group2026 — when longevity became a consumer category before an evidentiary oneDated 2026-09-01
Originated here
Biotech claims are made without denominators, and a dated, graded, revisable register of what the evidence actually says is a product rather than a publication.

Basis · Three properties built, 30 pages. h3alth.capital is not registered and the other two link to it. The register holds no entries and says so.

The Agency Layer2026 — the studio’s position on Web 4’s fourth conditionDated 2026-09-28
Originated here
Web 4’s four conditions converge only if an agency layer exists — the open protocols through which agents discover, identify, authorise, trust, transact with and audit one another — and that layer, not any single agent, is the durable product the studio is really building beneath every venture.

Basis · Every layer on the map is draft: github.com/FlashyLabs/web4 states the nine-layer stack as of 2026-09-28 with `delegation/1`, `intent/1`, `trust/1` and the rest all at draft and no independent adopter named. The layer is specified, not yet adopted.

Portable Authority2026 — the identity question the agent workforce forcesDated 2026-09-28
Originated here
Identity for agents is a chain, not a login: a workforce of software must prove not just who it is but who authorised it and to what extent, and the only credential that answers that is an attenuated grant chain a counterparty can verify without trusting the issuer.

Basis · Draft and unadopted: github.com/FlashyLabs/flashyid-spec carries `delegation/1` as a public draft on the 2026-09-28 map with the implementation private and no independent relying party named. Specified and enforceable, not yet used outside the estate.

Declared Existence2026 — the graph question underneath discoveryDated 2026-09-28
Originated here
The graph of what exists will be declared by owners, not scraped: an index an organisation publishes about itself, with every edge citing a source and none self-attested, is more accurate and more durable than any index a third party assembles by crawling.

Basis · Draft and unadopted: github.com/FlashyLabs/agentgraph lists `graph/1` public and draft on the 2026-09-28 map, built out from the private therealm registry, with no independent organisation serving a fragment. Specified, not yet populated from outside.

Published Intent2026 — the future tense of the record, made a formatDated 2026-09-28
Originated here
Organisations will publish intent as a file because meetings do not scale to agents: the past tense is already published everywhere and the future tense almost nowhere, so a decaying, machine-readable record of what a party wants to do next is the missing half of coordination.

Basis · Draft, with the reference implementation IntentMesh public and every emitted file still inside the estate: github.com/FlashyLabs/intent-spec carries `intent/1` public and draft on the 2026-09-28 map, and adoption-signal.json reads independentAdopters 0.

Consented Trust2026 — trust as routing rather than as a scoreDated 2026-09-28
Originated here
Trust routing beats reputation scores because a relationship cannot be bought: an introduction that crosses each edge only when its owner consents routes on the one asset activity cannot manufacture, where a score is a number and a number is always, eventually, for sale.

Basis · Draft and private: github.com/FlashyLabs/web4 lists the trust layer’s `trust/1` and `ritual/1` with Magician and Rites-Network both private on the 2026-09-28 map, and Magician’s four packages are on no registry. Built, not yet adopted outside the estate.

The Machine Door2026 — the gateway every site will need for agentsDated 2026-09-28
Originated here
Every site will need a machine door — a standard surface telling a visiting agent who to speak to and what may be asked — and the gateway that runs those doors is the CDN of Web 4: a thing everyone needs, cheap to run and expensive to be without.

Basis · Draft, open, unadopted: github.com/FlashyLabs/agent-wellknown carries `agent/1` public and draft on the 2026-09-28 map alongside `agent-dns/1`, with the Bastion service private-intended and no independent site serving the door. Specified, not yet a standard.

Chartered Organisations2026 — governance for a workforce that is softwareDated 2026-09-28
Originated here
The charter is the cap table of the agent workforce: a machine-readable declaration of an organisation’s roles, what each may do and who is accountable above it, is the governance record everything else — discovery, trust, spend, audit — derives from, and it belongs to the organisation rather than to a platform.

Basis · Draft and private: github.com/FlashyLabs/web4 lists `aao/0.1` for the representation layer as private on the 2026-09-28 map, with the estate’s own charters conformant but no independent organisation publishing one. The manifest and its checker exist; adoption does not yet.

The Mesh2026 — execution across organisations, not inside oneDated 2026-09-28
Originated here
Execution across organisations needs a shared place to agree, work and both hold the record: coordination between organisations is a different product from orchestration inside one, and the mesh is that place — where a human at each party consents before anything becomes a fact.

Basis · Draft, with the estate its own first tenant: github.com/FlashyLabs/flashyos-spec is public and draft on the 2026-09-28 map while the FlashyOS implementation is private, and adoption-signal.json records independentAdopters 0. Live for the estate, not yet outside it.

Policy-Bound Value2026 — how value moves when the spender is softwareDated 2026-09-28
Originated here
Value moves under human-set policy or it does not move: an agent may spend automatically inside a bound a person set — amount, purpose, expiry, counterparty — and nothing outside it moves at any speed, with every delegated policy only ever narrower than the one it inherited.

Basis · Draft, wallet and rails private: github.com/FlashyLabs/agentpay carries `pay-policy/1` as a public draft on the 2026-09-28 map over a private wallet and rails stack, with no independent counterparty settling against it. Specified and enforced, not yet adopted outside the estate.

Sealed Record2026 — the audit layer as a public utilityDated 2026-09-28
Originated here
A record one party holds is a claim, so the audit layer is a public utility: an append-only, hash-chained account any third party can verify without trusting whoever wrote it is the difference between evidence accepted in a dispute and a vendor’s log dismissed as self-serving.

Basis · Planned and private: github.com/FlashyLabs/web4 lists `action-ledger/1` as planned for the audit layer, generalising the private Chronicle pattern, with flashy-network the private notary consumer on the 2026-09-28 map. The pattern runs in the estate; the general format is not yet a repo.

Fewer Businesses Than Brands2026 — the studio’s own structural claim about itselfDated 2026-09-28
Originated here
The estate is roughly seven businesses under many more names, and the studio’s portfolio is the stack: a brand is a domain, a spec and a page, while a business is a licence entry, a support obligation and a person accountable — so the durable things are the protocol layers, not the logos over them.

Basis · Stated on the map and held by the register: github.com/FlashyLabs/web4 counts roughly seven protocol businesses plus a graph as of 2026-09-28, and the estate licence register in flashyos holds one entry per repository — a brand with no register entry is a name, not a business.

§ 05

M1 · Pre-thesis

A real observation about the world. Not yet a company, and named here rather than dressed as one.

FinanceOS2021 — the DeFi-as-a-service periodDated 2026-08-26
InheritedIn formation
Financial primitives would become a service layer other companies rent rather than a product end users choose, and the durable position is the rails rather than the interface.

Basis · Absent from every live register: no deploy/1 target, no directory fragment, no shipped/1 log. This is the origin of the sequence and it is not a running company.

rare.fun2025 — after the correction removed the speculative floorDated 2026-08-26
Originated hereIn formation
Scarcity is only interesting where it is enforceable, and consumer products are the fastest way to find out whether a scarcity mechanism is believed.

Basis · Status `in-formation` in portfolio.ts. Absent from every live register: no .deploy/config.json target, no directory fragment, no shipped/1 log. A written position, named here rather than dressed as a company.

§ 06

The register in date order

Each thesis carries the earliest artefact that could only exist if the belief already did. Read down the column and the descent is visible: the ledger before the settlement layer, the settlement layer before the coordination layer, identity after coordination needed it. Five entries are dated to the year because that is the resolution their record supports.

DatedVentureWhat dates itGrade
2021 · yrFlashy GroupThe first of the six transactions, published tombstone by tombstone. gda.group dates its record to the year, so this does too.M3
2021 · yrMetaverse GroupAcquired by Tokens.com, a listed acquirer, in a GDA-led sale. The record is dated to the year.M4
2021 · yrNFT BAZLThe inaugural convening, held while the category had money and no shared definition.M3
2025 · yrBitMap HoldingsAcquired by Skrybit in a GDA-led transaction.M4
2026-05-01ClaimYour.Goldb8248c78b, `Initial Commit` on the production line — 2,596 commits back, and the oldest dated artefact in the register.M3
2026-07-09Flashy Academy97faf8a, `Scaffold flashy-academy as a standalone Next.js app` — the earliest dated artefact of any venture in this register except ClaimYour.Gold.M3
2026-08-11Flashy Goldeb479fa, `Build flashy.gold site from the v2 mockup` — the first commit of the property.M3
2026-08-14Flashy Network6b93676, `Create flashy.network — the settlement record`. The commit subject is the thesis.M3
2026-08-18Flashy ID95d9f01, `Flashy ID: OAuth 2.1 / OIDC identity provider for inter-organizational agent coordination`.M3
2026-08-23BuildOS3f20b54, `Build OS landing page + pnpm lockfile` — a landing page for a property whose domain has never resolved.M2
2026-08-26FinanceOSNo artefact of its own. First named in 4.ventures' portfolio register, commit b1ef5d3, on 2026-08-26 — which is the honest anchor for a thesis that never became a company, and is why it grades M1.M1
2026-08-26FlashyOSac93495c, the flashyID identity foundation — the first commit of the mesh, and it is an identity commit, which is the thesis stated in the order the work actually took.M3
2026-08-26Flashy MindNo artefact of its own. First named in 4.ventures' portfolio register, commit b1ef5d3 — the same day as FinanceOS and rare.fun, and the three ventures with no independent record are exactly the three graded lowest.M2
2026-08-26rare.funNo artefact of its own. First named in 4.ventures' portfolio register, commit b1ef5d3.M1
2026-08-31IntentMesh1c60dc73, when the section got its own cards, addresses and machine-quotable claims — the day it stopped being a paragraph on somebody else's page.M2
2026-09-01Magiciane27d29f, `Harden the door before it is on fifteen properties, and pin the deploy`.M2
2026-09-01NFT Capital5811494, `An incubator, and a command that lets a project leave it` — the commit that created the incubating directory both this and h3alth were built in.M2
2026-09-01h3alth Group5811494, the same incubator commit as NFT Capital. Two brands staged in one afternoon is a fact about the studio's throughput, not about either brand's demand.M2
2026-09-02AssetMesh4a76b65a, `AssetMesh: rwa/1, the denomination layer for real-world assets` — the format landed before the pitch, which is the order this estate builds in.M2
2026-09-02RitualOS3b86dbed, `ritual/1: the present tense — a format for witnessed observance`. Same day as AssetMesh: two formats, one afternoon, neither with an adopter.M2
2026-09-22The wallet authorization planec35d7b40, PR #137 “Reconcile the wallet plane and Flashy Mind onto main (Phase 33)” — 619 files changed, +63,027/−151, merged onto main.M2
2026-09-22Countersign4964045 in flashy-network, ‘notary/1: fold backlog/1’s promoted items in as a fifth register’ — backlogToFragment turns each property’s promoted backlog items into intent/1 leaves in the transparency log.M3
2026-09-28The Agency LayerThe Web 4 stack map at github.com/FlashyLabs/web4, published 2026-09-28: the nine layers of the agency layer named, each with a per-repository status and an honest “no adopters yet”.M2
2026-09-28Portable AuthorityThe identity layer on the Web 4 map: github.com/FlashyLabs/flashyid-spec, contract `delegation/1`, listed public and draft on the 2026-09-28 stack map, with the implementation held private.M2
2026-09-28Declared ExistenceThe graph layer on the Web 4 map: github.com/FlashyLabs/agentgraph, contract `graph/1`, public and draft on the 2026-09-28 stack map — every edge citing a source, no self-attestation.M2
2026-09-28Published IntentThe discovery layer on the Web 4 map: github.com/FlashyLabs/intent-spec, contract `intent/1`, public and draft on the 2026-09-28 stack map — visibility has no override, expiry is computed, never input.M2
2026-09-28Consented TrustThe trust layer on the Web 4 map: github.com/FlashyLabs/web4 names `trust/1` and `ritual/1`, with routing (Magician) and standing (Rites-Network) both held private on the 2026-09-28 stack map.M2
2026-09-28The Machine DoorThe gateway layer on the Web 4 map: github.com/FlashyLabs/agent-wellknown, contract `agent/1`, public and draft on the 2026-09-28 stack map, serving a well-known agent surface; the commercial gateway (Bastion) is held apart.M2
2026-09-28Chartered OrganisationsThe representation layer on the Web 4 map: github.com/FlashyLabs/web4 names `aao/0.1` — the AAO manifest and its seven-question conformance — held private on the 2026-09-28 stack map.M2
2026-09-28The MeshThe execution layer on the Web 4 map: github.com/FlashyLabs/flashyos-spec, public on the 2026-09-28 stack map, with the FlashyOS implementation held private — agents suggest, humans consent.M2
2026-09-28Policy-Bound ValueThe commerce layer on the Web 4 map: github.com/FlashyLabs/agentpay, contract `pay-policy/1`, public and draft on the 2026-09-28 stack map — money in Minor integer units, a child policy only narrows, a human gate mandatory.M2
2026-09-28Sealed RecordThe audit layer on the Web 4 map: github.com/FlashyLabs/web4 names `action-ledger/1` (planned) generalising Chronicle’s sealed-receipt pattern, with the notary consumer flashy-network private on the 2026-09-28 stack map.M2
2026-09-28Fewer Businesses Than BrandsThe Web 4 stack map at github.com/FlashyLabs/web4, published 2026-09-28: roughly seven protocol businesses plus a universal graph, under more names than that — the layer is a business, a repo is a name.M2
§ 07

Originated by the studio

Ideated here, built here. Each entry states its status without decoration — including the ones that are currently nothing more than a written position.

BuildOSVertical AI · construction

Build

The operating system for construction firms: itemised quotes generated from job-site photographs and video walkthroughs, live project profit-and-loss, an automatic case-study and content engine, referral and loyalty programmes, inventory, and a marketplace for overflow leads.
Consumes 02Consumes 03Consumes 04
FinanceOSVertical AI · financial operations

In formation

The second application of the BuildOS pattern: an agent-operated operating substrate for a professional services vertical, with the same four elements — the work performed by agents, the customer relationship carried by portable identity, the incentives settled on a ledger, and the platform positioned to consolidate the firms that run on it.
Consumes 02Consumes 03Consumes 04
rare.funDigital collectibles · consumer

In formation

A consumer position on scarce digital objects, taken with the benefit of having operated the previous cycle’s version of the same idea and watched it reprice.
Consumes 02Consumes 03
AlchemistConsumer · play

Operating

The Bench: a client-side alchemy world built on the alchemy/1 format — seven operations over a world that is a document, where nothing is submitted and nobody is recorded. The play runs entirely in the reader’s browser.
Flashy BuzzConsumer · music & audio

Operating

The music and audio surface of the For-Gold economy: creator and fandom mechanics — sing-to-earn, tokenised fandom, audio creator economies — that pay out in the same rewards ledger the rest of the Flashy properties settle against.
Flashy WalletWallets · agent and consumer spend

Build

The wallet engine for agents and consumers: an object that can spend inside a limit a human set, built on Tether’s WDK toolkit, with the policy-guard and staking primitives open-sourced so a counterparty can check what an agent is allowed to do.
BastionInfrastructure · machine gateway

Design

The gateway that exposes an existing site or API as a secure, authenticated, payable machine interface — so a counterparty’s agent can transact against a surface built for humans, without the owner rebuilding it.
agentfileTools · organisational accountability

In formation

A writer over the published mesh formats that answers the accountability question about an organisation — who may act for it, with what authority, and what it has sealed — in about ninety seconds.

The one rule. Everything claimed here links to a live surface, a public API, or a commit that can be verified directly. Ambition is labelled North Star. When the work catches up, we update the page — not the other way around.

§ 08

What every venture page has to record

Every venture is described by which conditions it supplies and which it consumes, and by a thesis that could have been wrong. The studio cannot describe a company without stating both, and neither should a reader accept one.

FieldWhat it records
BeliefThe one sentence a partner would have to underwrite
Why thenWhat changed that made it possible in that year and not three years earlier
WedgeThe narrow first thing. Without one it is a market map
FalsifierOne observation that would kill it. One, not five
Grade and basisThe maturity, and the register it was read from
Condition suppliedThe one the founding team builds — the genuine novelty
Conditions consumedWhich group systems it runs on, and at what status
Live surfaceA URL, an API or a commit. Required.
Mesh organizationIts FlashyOS identity, because every venture runs on the mesh
§ 09

The intersections we are originating against

IntersectionWhat it produces
Agents × worldsPersistent-world labour — venues that stay open, operators that trade overnight, characters with real economies behind them
Agents × valueThe machine economy — agents earning, spending and settling in units that redeem
Agents × identityDelegated authority — an agent that can prove who authorized it and what it may do
Worlds × valueReal-world-asset backing for virtual property and in-world goods
Worlds × identityDigital citizenship — standing that persists across venues and surfaces
Identity × valueEarned standing convertible into real-world claims

Build one with us