Open Source opened software. Purpose Source opens its impact.

The thesis, at length

Open code. Shared progress. Public good. Purpose Source keeps what people love about community-developed software — public source, collaboration, forks, learning, reputation, free access for individuals and small organisations — and adds something open source does not create by itself: when sufficiently large companies benefit, a small licence fee is built to create real-world charitable impact. For the community, almost nothing changes. For the maintainer, the repository operates as before. The value the software already creates is simply allowed to travel further.

Don’t let useful code stop at being useful. Maintainers already give extraordinary work to the world — because knowledge should be shared, because software gets better together, because they want their work to be useful beyond themselves. Purpose Source asks: if that is already why you give your work to the world, why stop its impact at software? Connect the repository once; from then on the licence is built to turn the project’s continued existence, growth and use into public benefit. No fundraising campaigns. No asking contributors for money. No charging the community. The project simply gains another consequence.

Your accumulated work can begin doing good today. Years of development, thousands of commits, hundreds of contributors may already exist in a repository. Nothing needs rewriting; the community does not start again. A mature project that quietly served developers for a decade can start generating money for health, education or humanitarian causes simply because large organisations keep using its new releases. Precision, always: past releases stay under their old licence forever; the conversion aims the future.

Your code can help people who will never clone your repository. Your software can do more than run. It can have purpose.

Give contributors more than a green square. Open collaboration already pays in reputation. Purpose Source is built to add a measurable layer on top of it: attributed Impact Shares under a published algorithm, chosen causes, verifiable certificates. The contributor donates nothing — they write software, and it is the software’s usefulness that the licence turns into impact. Skill, not wealth, becomes the currency of philanthropy: a developer with no money to give can nevertheless direct funds to causes across a career, because their skill created the economic value.

Agency over the good. Contributors do not just learn that something good happened — they take part, choosing among verified cause categories. A contribution acquires a direction, and the better their work makes the software, the more potential impact that work can create.

Tiny cost per payer. Distributed impact by design. The Purpose Fee is deliberately not another expensive enterprise bill; it is priced to be the easiest compliant yes in the queue. For a large organisation the Pass is deliberately boring: one flat annual credential, cheaper than the meeting held to discuss it. One company paying a small fee does little; thousands of companies across a large ecosystem, year after year, is something else entirely. The asymmetry is the idea.

Something positive in return for companies. Large users are participants, not enemies: a clear commercial licence, central licence management, compliance records, invoices, a verified supporter profile, annual impact statements, certificates naming chosen cause categories. Licensing stops being purely transactional — within a published claim-language kit, and only once the ledger carries the row that backs the sentence.

A cross-project identity. Contributions today fragment across repositories: a hundred merged pull requests in ten projects leave ten unrelated histories. The design intent is one identity that spans them — a contributor profile that aggregates their impact across every registered repository they touch, a maintainer profile that aggregates their projects, a company account that aggregates its entitlements — so that the network connecting code, contributors, companies and public benefit is visible as one thing rather than inferred from many. None of it exists yet: the claim flow opens with platform v1, and the profiles carry figures only once ledger rows exist to carry.

Ownership stays where it belongs. The repository remains the maintainer’s: what merges, how it develops, which releases ship, who receives a discretionary waiver. The steward provides infrastructure; the maintainer provides the project; the community provides collaboration; large users provide the fee; public-benefit organisations receive the impact.

The central idea. The developer still writes code. The maintainer still maintains. The student still learns. The small business still builds free. But when an organisation large enough that the fee is trivial benefits from the collective work, the licence is built to route a fraction of that value out of the software industry entirely and towards the wider world.

The category

Purpose Source is a licensing category, not a company and not a foundation-branded programme. A project adopts it by committing one licence file. From that day, organisations large enough that the fee is trivial pay an annual Purpose Fee to keep using new releases, and the net proceeds route to vetted public-benefit funds under a published, capped, audited cost line.

Everything a community does with the code stays the same: public source, forks, pull requests, free use by individuals, students, nonprofits, and organisations below the threshold. No copyright assignment, ever — contributors keep their copyright. Each version becomes plain Apache-2.0 four years after its release, and that promise lives in the licence text rather than in a blog post.

Open Source opened software. Purpose Source opens its impact. That is a continuity claim, not a competitive one: open source needs no replacement, and this is not open source. We say so proudly, in those words, on the comparison page and here.

Adoptable without permission

Nobody has to ask us. Nobody hands anything over. The licence text is the product; this organisation is infrastructure around it — a registry, a payment rail, a ledger, a certificate authority for claims about who paid what. A project can adopt the licence and never speak to us. A project can leave the same way it arrived: change the file back. Past releases keep the licence they shipped under, forever, and every payer keeps every version whose publication date falls on or before the end of their paid term.

That decentralisation is deliberate and it is a constraint on us, not a feature we market. A movement that can only work through one organisation’s servers has recreated the problem it claims to solve.

What we will never be able to do

The steward’s statutes fix a small number of clauses that cannot be reopened by any future board: operating costs are capped, published, and audited; no distributable private profit exists — structurally, not by policy; the charity intermediary is curated rather than ad hoc; fee waivers for named organisations are gratis and public; the allocation ledger is append-only; and dissolution routes remaining funds to public benefit, never to members.

Neither the steward nor any repository owner can be a recipient of routed funds. This is the part we expect to be tested hardest, so it is written as structure and published as a document rather than asserted as a value.

Kill criteria, published in advance

We may fail. The honest form of that sentence is to publish, before launch, the conditions under which we would stop — and to make stopping safe for everyone who paid.

A payer’s worst case is capped in the licence text itself: versions already vested stay usable forever, and the four-year conversion continues to run whatever happens to this organisation. If the steward lapses, a backstop in the text turns the licence permissive rather than leaving adopters stranded with an unreachable counterparty.

The pre-registered criteria, the payer-held-whole guarantee, and the wind-down sequence live on the kill protocol page. They are a standing page, not an announcement we would have to be shamed into writing.

What we are asking

Of maintainers: one committed file, and the willingness to tell your community honestly what it changes and what it costs — including that some employers bar contributions to non-OSI-approved licences, which is a real cost we do not claim is zero.

Of large organisations: compliance, not generosity. One flat annual credential, machine verifiable, with a 60-day cure window and amnesty covenants on purchase.

Of ourselves: that nothing is claimed before it is real. Until money has moved, this site shows mechanisms and design intent, and every figure on it is labelled illustrative. The first impact number appears when the first ledger row does.

Operating principles

The constitution seed: eight principles the statutes are written to hold. The fourth is the one we expect to be tested hardest, so it is written as structure and published as a document rather than asserted as a value.

1. Source in public; community participation free.
2. Access without wealth barriers below the published threshold.
3. Large-scale commercial use creates public benefit — one published schedule for everyone.
4. No distributable private profit — structurally, not by policy: 100% of net proceeds after published, capped operating costs (cap: set with counsel before launch, audited). The cap is recorded in the constitution. statute clause
5. Radical transparency: an append-only public ledger, published algorithms, public waivers, a public schedule, and the fee stack published end to end.
6. The steward is registrar and router, never owner: repository admins hold administrative control and represent sufficient authority for their repositories; copyright stays with the rightsholders; waivers are the admins’, gratis, forever.
7. Contributors are recognised, never paid; their routing choice is honoured as an advisory designation.
8. The movement never claims to be open source; it claims to be worth existing anyway.

The pre-registered kill criteria, the payer-held-whole guarantee and the wind-down sequence are a standing page: the kill protocol. The clauses no future board can reopen are listed under what we can never do.