This window's qualifying projects
All figures from Open Collective public budgets, June 2026
Four OSS projects past $1k — a formatter that nearly ran dry, a browser engine in exile, a perf initiative, and a French PDF library
Prettier ($5,382/mo) nearly shut down in 2024 — it recovered only after a podcast disclosed the near-death. Servo ($7,150/mo, +62.5% YoY) shows what a Linux Foundation Europe fiscal host unlocks for a Mozilla-abandoned Rust engine. e18e ($2,094/mo, 798 stars) collects from Google and redistributes — stars don't matter, institutional alignment does. CourtBouillon/WeasyPrint (~$1,056/mo) is the floor case: a defensible niche, two French maintainers, and a stable European B2B user base that has no alternative.
The projects in this window span a 7x earnings gap: Servo clears $7,150/month on the high end, CourtBouillon just edges past $1,056/month at the floor. What they share is a publicly auditable number and no VC involved. Each one runs a materially different playbook.
통계 카드를 불러오는 중…
Prettier — $5,382/month, no dual license, almost went broke in 2024
What it is
Prettier is the opinionated JavaScript/TypeScript code formatter that has done more to end tabs-vs-spaces arguments than any ESLint rule. 52,100 GitHub stars, 4,700 forks, ~37 million npm weekly downloads, MIT licensed. 1 Used by roughly 60% of all JavaScript developers. 2
Publicly disclosed monthly earnings
~$5,382/month — estimated annual budget of $64,581, current treasury balance $65,627, total raised $256,577, total disbursed $190,950 (74.4% payout rate). 3 The money goes to two independent maintainers: Fisker Cheung receives $2,750/month for development and maintenance (his May 2026 payment was made in June 2026). 4 Sosuke Suzuki received consistent payments through at least May 2025; his current payment status is not confirmed from visible expense records.
Meta Open Source is the single largest sponsor at $49,000 cumulative. 3 This trips the "is it corporate-backed?" question. The answer is structural: Christopher Chedeau (Prettier's co-creator, Meta employee) is one of five Open Collective admins, receives no payments, and has no ownership over the fund. The four other admins — including contributors from webpack/swc, Excalidraw, and the broader JS ecosystem — are independent. Neither Fisker Cheung nor Sosuke Suzuki work for Meta. 4 Meta is a donor, not a controller.
The trigger: a 2024 near-death experience and a podcast appeal
The original unlock was a January 2022 blog post when the fund hit $50,000. Christopher Chedeau announced they would pay two maintainers $1,500/month each. His framing was honest: "$50,000 is a sizable sum but not enough for someone with the skills to work on it to do it full time. But at the same time we shouldn't let it sit unused." 2
The more instructive moment came in July 2024. On the Syntax.fm podcast (episode #794), Chedeau disclosed that the fund had burned through roughly $130,000 and had only ~$10,000 left — about three months of runway. 5 The episode also launched a t-shirt collaboration with $10 per shirt going to Prettier. By June 2026, the balance had recovered to $65,627 — a 6.5x increase in under two years. The November 2023 Rust bounty episode (where Biome claimed a $22,500 challenge) had already put Prettier's name back on developer timelines. 6 Corporate sponsors including Indeed ($20,000), Cybozu ($15,136), and JetBrains ($15,000) joined after those events. 3
The monetization model
Pure tiered sponsorship. No paid version, no dual license, no SaaS. Every dollar comes from voluntary donations via Open Collective (primary), GitHub Sponsors (49 current sponsors, $27,312 raised since February 2024), Tidelift, and a Drips Ethereum channel. 7 The GitHub Sponsors page states plainly: "We use donations to pay two maintainers $1,500 every month and one-off expenses." 7
Would this work for you?
Prettier's model depends on ubiquity and a clear "used in CI by every serious JS project" story that enterprises can verify. The fund burned to near-zero despite that ubiquity — and recovered only after a public near-death disclosure that generated fresh urgency. If your tool is widely embedded but your users don't feel the risk of it disappearing, the Prettier arc suggests that making that risk legible is the actual funding lever. The podcast moment did more than the $50k milestone post.
Servo — $7,150/month, 62.5% YoY growth, a browser engine that Mozilla dropped
What it is
Servo is a web browser rendering engine written in Rust, originally created by Mozilla Research in 2012. In 2020 Mozilla shut down the Servo team. In 2023 the project was accepted under Linux Foundation Europe and resumed active development. 37,100 GitHub stars, 3,700 forks, MPL-2.0 licensed. Latest release: v0.2.0 (May 31, 2026). 8 Its current positioning is as an embeddable rendering engine — not a standalone browser, but the component you'd use to add web rendering to an application or robotics system.
콘텐츠 카드를 불러오는 중…
Publicly disclosed monthly earnings
~$7,150/month — estimated annual budget of $85,804, current balance $72,906, total raised $127,094, total disbursed $54,189. Disbursement rate 42.6%, which is lower than Prettier's but consistent with a project still accumulating reserves for larger contracted work. 9
The trajectory matters more than the snapshot. In 2024 the project raised $36,570. In 2025 it raised $59,430 — a 62.5% year-over-year increase. Admin Delan Azabani's 2025 review post put it directly: "2025 was the second year of the Servo Collective, and in that year we raised 59,430.02 USD (+62.5% over 2024). Your donations have helped the Servo project grow with its community in a variety of ways." 10
The trigger: formal governance after corporate abandonment
The Linux Foundation Europe affiliation is the unlock here, but not for the obvious reason. It wasn't the LF brand that brought in sponsors — it was what formalization signaled: this project has a Technical Steering Committee, public monthly calls, a transparent funding request process (tracked publicly in servo/project#187), and a fiscal host that can handle vendor payments. 9 Enterprises that wanted to fund Servo before couldn't route a check through a random GitHub account. They can route it through a Linux Foundation Europe-affiliated Open Collective.
2025 expenses show the model in practice: Outreachy internship ($8,000), CI server infrastructure ($3,915), contributor experience work by Josh Matthews ($13,125). 10 These are real disbursements for real work — not just a trophy balance sheet.
The monetization model
Tiered sponsorship via Open Collective plus GitHub Sponsors, under Linux Foundation Europe fiscal hosting. No commercial product, no paid support tier, MPL-2.0 license (with copyleft implications that make dual-licensing less relevant for an engine). 8
Would this work for you?
The Servo pattern is specific to a project that (a) got dropped by a major corporation, (b) has clear technical uniqueness with no independent replacement, and (c) can absorb the governance overhead of a formal nonprofit umbrella. If your project fits (b) and (c), Servo shows that the formal affiliation isn't just bureaucracy — it changes the category of organization that can write you a check. Without it, you're limited to individuals and small companies. With it, you can reach procurement departments.
e18e — $2,094/month for cleaning up the JavaScript ecosystem's dependency debt
What it is
e18e (Ecosystem Performance) is a community initiative — not a single library or tool — focused on improving performance across the JavaScript package ecosystem. Three workstreams: Cleanup (modernizing dependency trees), Speedup (optimizing tools and libraries), and Levelup (building lightweight modern alternatives). 798 GitHub stars as of June 2026. 11 Admin: James Garbutt, independent maintainer.
Publicly disclosed monthly earnings
~$2,094/month — estimated annual budget of $25,131, current balance $11,733, total raised $19,840, total disbursed $8,107. 12 The project has a stated goal of $7,000/month for "part-time team lead" — so it is currently at 30% of that target.
The donor composition is distinctive. Chrome's Web Framework & Tools Performance Fund (Google) contributed $10,000. Individual donor Theo Browne put in $5,000. Vitest contributed $1,800. Anthony Fu Fund $800. 12 Two of those four are entities with a direct performance interest in JavaScript tooling speed.
The trigger: Google's explicit bet on ecosystem performance
e18e crossed $1k/month primarily because Google's Chrome team has a budget category for exactly this — tooling performance that affects their own benchmarks and developer experience data. The fund's stated use is "sponsoring people within the community, sponsoring larger ongoing projects/work, community events." 12 That's not a donation model; it's a sub-grant model. e18e collects from large upstream beneficiaries and re-distributes to contributors doing the actual work.
The monetization model
Community initiative sponsorship with corporate anchor donor (Google Chrome team). The 798-star count is not what's funding this — it's the alignment between e18e's work and what Google needs to happen in the JS ecosystem regardless.
Would this work for you?
The e18e model only works if your project sits at the intersection of what a large company needs to happen and what no one is paid to do full-time. Google needed someone to coordinate JS dependency cleanup; they didn't want to own it. e18e exists to fill that gap. If you can identify a large company with a budget for "ecosystem health" and a specific problem you can address credibly, the sub-grant model is reachable. It does not scale through more stars. It scales through identifying the right institutional anchor.
CourtBouillon — €11,729/year (~$1,056/month), two maintainers, WeasyPrint turning 15
What it is
CourtBouillon is a French open-source collective building Python document-generation tools. The flagship project is WeasyPrint — an HTML/CSS-to-PDF renderer used by companies that need server-side PDF generation without a browser dependency. The collective also maintains pydyf (low-level PDF generation), tinyhtml5 (HTML5 parser), and slight (slideshow editor). Admins: Guillaume Ayoub (@liZe) and Lucie Anglade. 13
Note on GitHub stars: the main WeasyPrint repository sits under the Kozea organization (the original creator group), not the CourtBouillon GitHub org. CourtBouillon's org repos include weasyprint-samples (403 stars) and pydyf (150 stars). 14
Publicly disclosed monthly earnings
~$1,056/month — estimated annual budget of €11,729 (~$12,667 USD), current balance €4,615, total raised €58,653, total disbursed €44,887. Disbursement rate 76.5% — almost identical to Prettier's 74.4%, meaning CourtBouillon at one-fifth the income is paying out at nearly the same rate. 13 Fiscal host: Open Source Europe. Goal: "Two Days per Week" at €24,000/year — currently at roughly 49% of that.
Top donors are European software and consulting companies: Spacinov (€13,600), Kobalt (€5,163), Simon Sapin (€3,406 individual), Grip Angebotssoftware (€3,260). No VC, no US tech giant in the top four. 13
In June 2026, Lucie Anglade posted that WeasyPrint has turned 15 years old. 13 At €11,729/year, that is 15 years of software that processes production documents at companies across Europe, sustained by two people and a handful of consulting firms.
The trigger: a niche with no adequate commercial alternative
WeasyPrint's funding didn't come from a viral moment. It came from being genuinely useful to a specific class of user — European businesses that need HTML-to-PDF pipelines, operate under stricter document compliance requirements than US counterparts, and can't rely on headless Chrome in production. None of the donors are household names. All of them need the tool to keep working.
A May 2026 update from Lucie Anglade noted: "April was a quite difficult month for us on the personal side and then, we unfortunately didn't have a lot of time to work on WeasyPrint." 13 That's the transparent maintainer communication style that makes niche sponsors keep paying. They know what they're paying for.
The monetization model
Tiered sponsorship with European consulting firm anchors. WeasyPrint v69 shipped in June 2026. Development cadence is sustainable at current funding — not accelerating, but not at risk of abandonment. 13
Would this work for you?
CourtBouillon is the floor case for this series. €11,729/year represents two part-time people doing honest maintenance work on a tool that solves a specific unsexy problem for businesses that depend on it. No viral bounty, no Google anchor donor, no foundation affiliation. Just a defensible niche, transparent communication, and a set of companies that calculated the tool is cheaper to sponsor than to replace. If you maintain something with a defined professional user base and no adequate alternative, this is an achievable steady state — not exciting, but sustainable.
Model-fit matrix (updated)
| Archetype | Tiered OC/GH sponsorship | Corporate anchor grant | Foundation affiliation | Nonprofit governance | Dual license |
|---|---|---|---|---|---|
| CLI tool / formatter | ✅ Prettier — 74% payout, recovers from near-zero with public urgency signal | ⚠️ Meta as large donor, not controller | ❌ Overhead usually exceeds benefit | ❌ Not applicable | ❌ MIT license; community would resist license change |
| Browser / rendering engine | ⚠️ Possible once trust established | ❌ No single corporate anchor | ✅ Servo — LFE affiliation unlocks enterprise procurement | ✅ Servo — TSC + public calls + transparent budget | ⚠️ MPL-2.0 already has copyleft; dual-license less relevant |
| Ecosystem initiative (no single artifact) | ⚠️ Low individual conversion without a concrete tool to point to | ✅ e18e — Google Chrome team as primary funder | ❌ Not needed at current scale | ❌ Not applicable | ❌ Not applicable |
| Niche document/data library | ✅ CourtBouillon — European consulting firms as anchor sponsors | ❌ No tech giant interest in HTML-to-PDF | ❌ Overhead too high | ❌ Not applicable | ⚠️ Possible; limited upside if user base is B2B already paying |
| Library (widely embedded) | ✅ core-js (prior issue) — corporate risk framing | ⚠️ Possible if library blocks a key supply chain | ⚠️ Only if spec-adjacent | ⚠️ JSON Schema model (prior issue) | ✅ If business users are the primary consumers |
| Full-stack generator / framework | ✅ JHipster (prior issue) — 82% payout | ✅ JHipster — Okta as $105k lifetime donor | ❌ Separate from nonprofit; LF path is different | ✅ JHipster — French association, 13-year track record | ⚠️ Possible for Apache-licensed stacks |
Four patterns now visible across eight issues of this series: (1) individual sponsorship alone plateaus around $3–4k/month for high-star projects — getting past that requires at least one institutional anchor; (2) making the risk of abandonment legible (a near-death post, a podcast, a governance milestone) moves more money than asking harder; (3) the projects with the highest disbursement rates — JHipster at 82%, Prettier at 74%, CourtBouillon at 76% — share one trait: the fund is treated as payroll, not a reserve; (4) niche beats scale when the niche is irreplaceable. Servo at 37k stars makes more per month than Eleventy at 20k because its failure would strand a specific category of application with nowhere to go.
Cover: AI-generated illustration.
참고 출처
- 1GitHub: prettier/prettier
- 2Prettier Blog: Prettier begins paying maintainers
- 3Open Collective: Prettier
- 4Open Collective: Prettier expenses
- 5Syntax.fm: Prettier JavaScript with Vjeux, Syntax #794 transcript
- 6Prettier Blog: $20k Bounty was Claimed
- 7GitHub Sponsors: @prettier
- 8GitHub: servo/servo
- 9Open Collective: Servo
- 10Open Collective: Servo Collective 2025 in review
- 11GitHub: e18e/e18e
- 12Open Collective: e18e
- 13Open Collective: CourtBouillon
- 14GitHub: CourtBouillon organization
이 콘텐츠를 둘러싼 관점이나 맥락을 계속 보강해 보세요.