Should Birdbrain give itself a legal body — and should it be a Wyoming UNA, then a DUNA?

Birdbrain already is an unincorporated association in the eyes of a US court — twenty-three people joined by mutual consent, voting on-chain, pursuing a common purpose — but it is the common-law kind, which carries all of the exposure and none of the protection, and the cheapest fix available is to register under Wyoming's Unincorporated Nonprofit Association Act, which costs five dollars and converts that same set of people into a legal entity whose members are not personally liable for its debts.

Imagine a band. Five people who agreed, mostly by playing together rather than by signing anything, that they are a band. They have a name, a shared account for gig money, and they make decisions by arguing until everyone shrugs. Nobody has ever registered anything.

Now the van they hired damages a venue. Who pays?

The uncomfortable answer, in most legal systems, is: all of them, personally, without limit. Not because anyone tricked them, but because the law has a default box for "group of people acting together with no registered form", and that box does not have a lid. Whatever the group owes, the members owe. The drummer's flat is on the table.

This is not a hypothetical for decentralised organisations. It is the actual holding in the two US cases that most closely resemble what we are. In 2023 a federal court entered judgment against Ooki DAO for $643,542 on the basis that it was an unincorporated association made up of the token holders who had voted.¹ In 2024 another federal court let a case proceed on the theory that Lido DAO was a general partnership and that its large token holders could be personally on the hook as partners.² The judge in the second case put it plainly enough: participating in governance is participating, and a claim to be decentralised is not a defence.³

Birdbrain has twenty-three members. Every one of them holds a membership item on Kusama through Kreivo, and the whole design of this project — what we call naive joining, where taking part is membership — makes the "voluntary group of persons joined by mutual consent for a common purpose" test almost embarrassingly easy for a regulator to satisfy. We built the evidence file ourselves, on a public chain, on purpose.

So this is not a question of should Birdbrain be an unincorporated association. It already is one. The question is which kind: the common-law kind that offers members nothing, or a statutory kind that offers them a shield.

Wyoming happens to sell exactly that shield, and the price is not a metaphor. The Wyoming Unincorporated Nonprofit Association Act makes the association a legal entity separate from its members, says members are not personally liable for its contracts or torts, lets it hold property and sue in its own name — and the only filing involved is an optional five-dollar statement naming someone to receive legal post.⁴⁵ Two members is the minimum. There is no certificate of formation, no annual report, no franchise tax.

There is a second, newer Wyoming form built specifically for organisations like ours — the Decentralized Unincorporated Nonprofit Association, or DUNA — which treats a smart contract as a governing document and gives the association perpetual existence, an administrator role, and a proper record-inspection regime. It needs one hundred members. We have twenty-three.

That gap is the shape of this whole document. The legal work is cheap and mostly thinking. The hard part is being able to look at a hundred members and say honestly what each of them did.

Everything named here is defined from scratch. Skip any box you already know.

Birdbrain

Birdbrain is an open-source project that turns a community's activity — its calls, writings, arguments and decisions — into a shared, queryable world model that the community owns. It is the proving ground where the machinery was built: on-chain membership, self-custody identity, the passkey surfaces, the editorial layer. Decent and Four To The Floor are brands standing on that machinery. Birdbrain is the foundation project, and the question of what body it has is therefore a question about the foundation, not about any one brand.

The membership, measured

Birdbrain's Kreivo community is 1786. A live read of the chain on 21 August 2026 returned 23 issued memberships across 23 distinct holders. Decent Partners' own community (11) holds 4. Four To The Floor (444) holds 0. There is no ambiguity to argue about here: twenty-three is the number, and a DUNA needs one hundred, with no waiver in the statute.

What "no legal body" actually costs us today

Four concrete things Birdbrain currently cannot do, all of which it will need to do:

1. Sign a contract in its own name. Today any agreement is signed by Decent Partners Ltd, which means Decent Partners carries the risk and, worse, appears to own the thing.
2. Hold the intellectual property — the Seeds codebase, the protocol specifications, the name. It sits with a UK company. If Birdbrain is genuinely a foundation project rather than a Decent product, that is the wrong place for it to sit, and everybody can see that it is.
3. Receive money as itself, whether a grant, a treasury disbursement, or the proceeds of an Initial Graph Offering.
4. Absorb liability. Right now there is nothing between a claim and the twenty-three of us.

The UNA — Wyoming Unincorporated Nonprofit Association Act

W.S. 17-22-101 to 17-22-115. Fifteen short sections, on the books since 1993, adopted from a uniform act, and completely uninterested in blockchains. What it requires is almost nothing: two or more members joined by mutual consent for a common nonprofit purpose. What it gives is substantial:

• The association is a legal entity separate from its members for determining and enforcing rights, duties and liabilities in contract and tort.
• Members are not personally liable for the association's contracts or torts, and the association's torts are not imputed to them.
• It can acquire, hold, encumber and transfer property in its own name, and can be a legatee, devisee or beneficiary.
• It can sue and be sued in its own name.
• A judgment against it is not by itself a judgment against a member.
• It may file a statement appointing an agent for service of process for five dollars.

Note the word "nonprofit". It does not mean nobody can earn anything. It means the association does not distribute profit to members as owners. Birdbrain stewarding an open protocol and a shared world model fits comfortably. It is not the same as "charitable", and that distinction matters later.

The DUNA — Wyoming Decentralized Unincorporated Nonprofit Association Act

W.S. 17-32-101 to 17-32-129, enacted as SF0050 in March 2024, in force 1 July 2024, amended by SF0022 in force 1 July 2026.⁶ Same species, crypto-native organs bolted on. To qualify: at least one hundred members joined by mutual consent under an agreement — which may be written or inferred from conduct — for a common nonprofit purpose; the association must elect to be formed under the Act; and it must not be formed under another organisational statute.

The load-bearing invention is "governing principles", defined to include agreements, DUNA agreements, consensus formation algorithms, smart contracts and enacted governance proposals, whether recorded, implied from established practice, or both. Preston Byrne's section-by-section reading is right that this is the genuinely unusual part: Wyoming simply declared that for DUNA purposes the smart contract is the contract, with no plain-language equivalent required.⁷ It would still be a bad idea to rely on that alone, and we would not.

The 2026 amendment did four things worth knowing. It conformed the two acts to each other and made the downgrade explicit — fall below 100 members and the DUNA automatically becomes a plain Wyoming UNA. It added a federal-law hook to winding up, so remaining assets must go where state or federal law directs, or as the governing principles say, and absent governing principles, pro rata to members. It added a duty of care for administrators winding up, and inserted a "charitable purpose" definition referencing IRC 501(c)(3) for merger purposes. And it defined "distributed ledger protocol", letting the governing principles specify whether the ledger is immutable or mutable, public or private.

Read that third point carefully. Several commercial write-ups claim a DUNA must have a charitable purpose. The statute does not say that. It requires a common nonprofit purpose, and the 501(c)(3) reference sits in the merger and winding-up sections. Birdbrain's purpose is comfortably nonprofit and not obviously charitable, so this is a place where a service provider could quietly draft us into an obligation we cannot meet.

The administrator question, which is the one that matters

Both acts contemplate someone who acts for the association. The obvious assumption is that this is a person at a desk. For Birdbrain it is not. The association's administrator would be the operator key for community 1786 — an on-chain execution identity, bound by the membership's votes, holding no discretion of its own. Strategic decisions are taken by the members, on-chain. The key executes.

This is not decoration. It is the structural claim on which the tax analysis, the liability analysis and the honesty of the whole thing rest, and it is the first thing counsel should be asked to stress-test.

Three bodies of evidence: what a large ecosystem did when it faced this exact question and chose the expensive answer, what a large protocol did when it chose the cheap answer, and what courts have done to organisations that chose no answer.

Polkadot: the Cayman route, in full

Birdbrain's membership lives on Kusama, so Polkadot's journey is the most directly instructive precedent available, and it is worth walking through properly rather than gesturing at.

The problem Polkadot had. For years the ecosystem's off-chain work — contracts, fiat payments, intellectual property, physical presence — was carried by the Web3 Foundation and Parity Technologies. Both are philosophically aligned with the network. Neither gives DOT holders any formal role or right, and, as the proposal itself conceded, neither is well placed to.⁸ As governance decentralised and those organisations began handing functions to the community, the gap became structural: OpenGov could decide anything and execute only what a chain can execute.

What was built. Community members engaged Actum Node LLC, a consultancy that had been working closely with Parity leadership for around two years, to design a legal entity that could act as a real-world extension of on-chain governance. Working with community members and legal advisers, they produced a full package — bylaws, articles of association, memorandum of association, named directors, a named supervisor, a selected administrator — before asking for a single vote. Referendum 730, on the Treasurer track, requested $600,000 (approximately 88,000 DOT at the seven-day average price of $6.82) to incorporate the Polkadot Community Foundation and run it for twelve months. Funds went to a transition-team multisig held by Autonomous Projects, Actum Node and Integral Labs, to be released to the Foundation once the community could verify it had been set up properly. It passed and executed in mid-2024.⁸

The structure. A Cayman Islands foundation company — and the specific features are the point:

• Memberless. No shareholders or members hold any economic claim over the Foundation. The proposal is explicit that this eliminates a common source of conflict between a legal entity and the community it serves.
• A supervisor, whose job is to oversee the board and enforce the governing documents, with legal standing to act if directors fail their fiduciary duties.
• Five directors: a three-director majority of Cayman-resident professional governance providers, chosen both for independence from Polkadot entities and to satisfy Cayman economic substance requirements, plus two seats reserved for Polkadot governance to fill.
• An outsourced administrator, Autonomous Projects, a Cayman-based professional services firm, chosen competitively.
• DOT-holder rights written into the bylaws: delegate activities by referendum, remove the supervisor or any director by vote on the General Admin track, amend the bylaws by vote, wind the Foundation up by vote, receive quarterly public transparency reports, and — the sharpest lever — simply stop funding it, since OpenGov treasury proposals are its only income.
• The treasury stays on-chain. The proposal draws this line deliberately: "Unlike many other blockchain foundations, the Polkadot treasury will remain on-chain under the direction of OpenGov, and will not be subsumed by the Foundation."

Why Cayman and not Switzerland. The FAQ answers this directly and unsentimentally: a Swiss foundation is a valid option, but Swiss foundations are subject to income tax by default unless authorities grant an exception with conditions attached, whereas Cayman foundation companies are tax-exempt by default. Operational ease and flexibility decided it.

Why not Kusama. This is the part of the story most relevant to us, and it is easy to miss. The proposal was asked whether the Foundation could serve Kusama as well, and declined: because the Foundation reserves rights and powers for token holders, "it is unclear how the combination of DOT and KSM holders, with differing governance instances, can jointly indicate their preferences." The suggestion was that a separate foundation for KSM holders might be created later, and that the pattern might eventually be extended to any parachain or DAO in the ecosystem.⁸

That later foundation does not exist. There is no ecosystem legal body serving Kusama-based communities. Birdbrain is a Kusama-based community. Nobody is coming to wrap us.

What it costs to run. The published twelve-month budget is worth reading as a price list for the memberless-foundation pattern:

Asked whether the budget could be cut, the proposers said it was close to the bare minimum for compliant, redundant, professional-quality governance, noting that some foundations run with a single director and no administrator but that this would not suit the Foundation's goals.

What happened next, honestly. The PCF has done real work — the Polkadot mobile app, the Marketing Bounty, a partnership with Politecnico di Milano's Blockchain & Web3 Observatory, contracting the Polkadot-API team, and holding intellectual property for the Polkadot App that nothing else in the ecosystem is positioned to hold.⁹ Referendum 1591 renewed its funding for a second year. A Hong Kong subsidiary followed.¹⁰ But three problems surfaced that we should learn from rather than rediscover:

• The two community director seats sat empty. They were provided for from the beginning, and, as the community's own write-up puts it, "no one initiated this process." The election only ran in late 2025, through Referendum 1749 — more than a year in which the seats designed to represent token holders inside the Foundation represented nobody.¹¹¹²
• Instruction latency is brutal. The Marketing Bounty proposal spells out the arithmetic: each instruction to the PCF needs its own OpenGov proposal (1 to 35 days depending on track), followed by a 60-day lockup period for PCF approval, then contract negotiation and legal review. That is a governance loop measured in quarters.¹³
• The protection is narrower than it looks. The same discussion concluded that Cayman-based protection for bounty members is limited, and that non-signatories can still face exposure in certain configurations — with some participants arguing a dedicated entity would serve better than routing through the PCF. The May 2025 OpenGov.Watch report was blunter still, describing inefficiency, high costs, paperwork and disengaged directors, while conceding there was no clear alternative holder for the App's IP if the Foundation lapsed.¹⁴

The lesson for Birdbrain. Polkadot's answer is the right answer to Polkadot's question. A network with a multi-billion-dollar treasury, thousands of governance participants across every jurisdiction, and real litigation exposure needs a professional, liability-absorbing, memberless wrapper with named fiduciaries, and $600,000 a year is a rounding error against what it protects. But look at what that structure costs in the currency we actually care about: it puts five humans with fiduciary duties between the members and the decision, it introduces a 60-day approval lockup, and it creates an entity that can, in principle, refuse an instruction the community voted for.

Birdbrain has twenty-three members, no treasury to speak of, and a design premise that the operator key executes the membership's votes without discretion. Buying Polkadot's answer would cost roughly twenty-six times our entire projected annual running cost, and it would install exactly the human intermediation layer that the naive-joining pattern exists to avoid.

The DUNA is the other end of the same design space: member-based instead of memberless, no directors required, the smart contract admitted as a governing document, and state fees measured in single-digit dollars.

Uniswap: the DUNA route

Uniswap governance adopted a Wyoming DUNA — DUNI — by on-chain vote in September 2025.¹⁵ The executed shape is a usable template:

• an Association Agreement, the written governing principles, ratified by on-chain vote,
• an Administrator (Cowrie, David Kerr's firm in Cheyenne) engaged for the IRS interface, tax filings and informational reporting, paid $75,000 in UNI,
• a Ministerial Agent (the Uniswap Foundation) authorised to execute governance decisions with no discretionary authority,
• a DUNI-owned multisig prefunded with $16.5m in UNI for back taxes and legal defence,
• membership defined as holding more than zero UNI and participating in governance.

Note the ministerial agent. It is the same idea as our operator key: a thing that executes and does not decide. Uniswap needed a legal person to fill that role. We have a cryptographic one, which is arguably cleaner, and is certainly cheaper.

The instructive part is what happened after. Cowrie's update to the Uniswap forum records that DUNI filed Form 1120 for prior periods, paid tax, interest and penalties, filed Forms 843 seeking penalty abatement, and is pursuing a private letter ruling from the IRS, because the corporate-tax election could not be made retrospectively via Form 8832 for those periods. That is an expensive lesson pointing in exactly one direction: decide the tax posture and make the election before the entity has a history, not after. Form 8832 has a 75-day window. It is the single easiest thing to get wrong.

Other live adopters: Syndicate Network Collective (2024) and Towns. Alabama has since passed its own DUNA act, so Wyoming is no longer the only domicile.

The courts: what "no answer" costs

CFTC v. Ooki DAO. In June 2023 Judge Orrick of the Northern District of California entered default judgment against Ooki DAO, awarding a $643,542 civil penalty and permanently enjoining its operations. The court found the CFTC had sufficiently pleaded that Ooki DAO was an unincorporated association under both California and federal law — a voluntary group of persons, without a charter, formed by mutual consent, to promote a common objective — and therefore a "person" capable of violating the Commodity Exchange Act. The CFTC's theory throughout was that the association comprised "Ooki Token holders who have voted those tokens to govern the Ooki Protocol."¹¹⁶

Read the four-part test again with Birdbrain in mind. Voluntary group of persons: yes. Without a charter: yes, today. Formed by mutual consent: yes, and we log the consent. Common objective: yes, and we publish it.

Samuels v. Lido DAO. In November 2024 Judge Chhabria, same district, denied a motion to dismiss on the theory that Lido DAO was a general partnership under California law and that large token holders participating in governance could be liable as general partners.² The reasoning turned on collective profit-making activity and the governance role of token holders. Commentary since has been consistent and uncomfortable: an unwrapped DAO structure appears to provide negligible liability protection, and a claim to decentralisation is not itself a defence.¹⁷³

Two distinctions matter for us, and they cut in opposite directions. In our favour: both cases involved profit-seeking organisations, and Birdbrain is nonprofit in purpose, which weakens the general-partnership analysis considerably. Against us: the Ooki holding is about unincorporated associations, which is precisely what we are, and a common-law unincorporated association generally does not limit its members' liability. Registering under W.S. 17-22 is the difference between the Ooki fact pattern and an entity with statutory limited liability. That is the argument for spending five dollars, in one sentence.

The ladder

Phase 0 — decide, before anything is drafted. Four decisions, all of which are thinking rather than filing. The purpose clause, which must be nonprofit and must not be accidentally charitable. The membership definition, which has to be one we can defend honestly. The tax posture, including whether to elect corporate treatment on Form 8832 within 75 days. And the arm's-length instrument that moves the intellectual property from Decent Partners Ltd to the association, which is the step that makes "Birdbrain is a foundation project, not a Decent product" true rather than merely stated.

Phase 1 — stand up the UNA. Founding members. An Association Agreement that binds the association to community 1786's referenda and names the operator key as administrator. A Wyoming registered agent (roughly $50–200 a year) and the $5 agent-for-service filing. An EIN on Form SS-4. The 8832 election. A treasury. Registry entries per our own billing and onboarding rules.

Phase 2 — operate it so it is real. Grow 23 towards 100 through the passkey surfaces. Keep a dated member register that reconciles to the chain. Run actual governance. File the 1120.

Phase 3 — elect into the DUNA, once the membership genuinely clears 100: amend the governing principles to elect under W.S. 17-32, and pull the smart-contract layer explicitly inside "governing principles".

Phase 4 — the Initial Graph Offering gate. Do not start this before securities counsel on both sides of the Atlantic.

What it costs

Set that against $600,000 a year for the Cayman pattern. The UNA phase is nearly free in state fees and can be done with a good Association Agreement plus a targeted review. What genuinely needs paying for is the tax election, the UK analysis, and the IP instrument — and those are worth paying for now rather than after a DUNI-style back-tax letter arrives.

The three things that could sink it

The UK question. Everyone here is British. HMRC's corporate residence test is central management and control: a body incorporated abroad is UK tax resident if the highest level of control over its real business abides in the UK, as a question of fact rather than of paperwork.¹⁸ The naive version of this risk assumes a human administrator deciding things from a desk in north London, which is the wrong model — the administrator is the operator key, the members vote, and no single member decides. That materially changes the factual basis HMRC would analyse. It does not make the question disappear. Get a UK opinion alongside the US one, and model it before anything is filed.

The FCA clock. The UK's cryptoasset regime commences 25 October 2027, and the authorisation application window runs 30 September 2026 to 28 February 2027.¹⁹ Firms relying on a section 21 approver that do not apply in that window must run off their UK business before commencement. If any IGO communication touches UK retail persons, we are inside that regime, with a window that opens in weeks. The answer may well be that IGO #1 simply does not promote to UK retail, but that is a decision to take deliberately and on a date.

The hundred-member wall. Twenty-three today. The temptation will be to reach a hundred by minting, and naive joining makes that trivially easy, which is exactly why it is dangerous. A member set that cannot survive the question "what did these hundred people actually do" is worse than no wrapper at all, because it converts a governance story into an evidence problem. Whatever membership definition we adopt has to come with a lapse rule, and the register has to reconcile to the chain on a date.

"Why do this at all? Nobody is suing us." Correct, and the honest framing is not fear of litigation. It is that Birdbrain cannot currently sign, hold, or receive anything as itself, so every commitment routes through a UK company and every observer can reasonably conclude Birdbrain is a Decent product. The liability point is real but secondary; the ownership point is the one that changes what this project is.

"Wyoming is a gimmick jurisdiction." It is a small state that legislates early for this sector, which is a fair criticism and also the reason the statute exists at all. The mitigation is that we are not relying on novelty: the UNA Act is a 1993 uniform-act adoption with three decades of ordinary use behind it, and the DUNA is an upgrade path, not the entry point. Alabama now offers an alternative, which is worth pricing.

"A Cayman foundation is the proven path — Polkadot did it." Polkadot did, for good reasons, at $600,000 a year, with five human directors and a 60-day approval lockup. Their own proposal declined to extend it to Kusama, so the path is not actually open to us on those terms. And the shape is wrong for us: memberless where we are member-based, discretionary where our operator key is deliberately ministerial.

"Doesn't a legal wrapper undermine the point of being on-chain?" It would if the wrapper made the decisions. The Uniswap design and ours both answer this the same way — the executing party has no discretionary authority — and the DUNA statute is unusual precisely because it admits the smart contract as a governing document rather than requiring a paper equivalent to sit above it.

"Twenty-three members voting on their own legal form is not much of a mandate." True. That is why this is an RFC before it is a referendum, why the endorsement threshold is 20 rather than a number we cannot reach, and why nothing gets filed in Wyoming before the membership has voted on the actual Association Agreement text.

"Should Decent Partners Ltd really give away the IP?" It is the load-bearing question and it is not free. The instrument has to be arm's-length, defensible to both HMRC and the IRS, and drafted by someone who does this. But a foundation project whose foundations are owned by one member's UK company is a foundation project in name only.

Open questions to the room:

1. What should a Birdbrain member actually be? Holding a 1786 membership item is the mechanical answer. Is it the right one, and what makes a membership lapse?
2. Does anyone object to the operator key as administrator — and can you break it? Specifically: what happens if the key is compromised, or if a vote is ambiguous, or if the chain is down when something must be signed?
3. UNA now, or wait and do the DUNA once? The case for now is that the shield and the IP transfer are worth having at 23 members. The case for waiting is one drafting exercise instead of two.
4. Wyoming or Alabama? Has anyone read Alabama's act closely enough to say whether it is materially different?
5. Is $600,000-a-year Cayman ever the right answer for us? If the IGO succeeds beyond expectations, does the calculus flip — and at what number?
6. Who should the counsel be? Candidates so far: Cowrie (Cheyenne — David Kerr co-drafted the statute and administers DUNI), Falcon Rappaport & Berkman, Preston Byrne / Byrne & Storm (an English lawyer first and an American second, which suits our transatlantic problem), DAObox for turnkey. UK side, Own Your Numbers for tax modelling plus a crypto-regulatory firm for the FCA question.

Primary statute. Wyoming Unincorporated Nonprofit Association Act, W.S. 17-22-101 to 17-22-115. Wyoming Decentralized Unincorporated Nonprofit Association Act, W.S. 17-32-101 to 17-32-129, as introduced (SF0050, 2024) and as amended (SF0022, Enrolled Act 21, 2026 Budget Session, in force 1 July 2026). The Secretary of State's own entity guide confirms the $5 fee and the absence of any formation certificate, annual report or franchise tax for a UNA.

The membership figure. 23 issued memberships across 23 distinct holders in Kreivo community 1786, read live from the chain on 21 August 2026 via /api/membership/aggregate on the decent-kreivo-api service. Anyone can re-derive it from the chain directly; if it has moved since publication, the later number wins.

The PCF figures are taken from Referendum 730's own text on Subsquare and from the Polkadot Wiki's PCF page, both linked below. The budget table is reproduced from the proposal verbatim. The $600,000 was requested as approximately 88,000 DOT using a seven-day average price of $6.82; the DOT figure therefore reflects mid-2024 prices and should not be read as a current cost.

What this document does not claim. It does not claim we have taken legal advice — we have not, and nothing here is a legal opinion. It does not claim the Ooki or Lido holdings apply directly to a nonprofit association; both involved profit-seeking organisations, which is a material difference that cuts in our favour. It does not claim a Wyoming registration solves the UK tax question, which is separate and unresolved. It does not claim the PCF has failed — it has delivered real work, and the criticisms recorded here are the ecosystem's own, not ours. And it does not claim we can reach 100 members honestly. That is the open question the whole plan rests on.

Corrections are actively wanted. This is a discussion draft, not a decision. If you have stood up a DUNA, sat on a foundation board, or been on the receiving end of a regulator's view of an unincorporated association, that is the most useful reply we could get.