marketplace-fee-data

Seller fees for ten online marketplaces, as a plain JSON file, plus a zero-dependency helper for the three questions everyone actually asks: what does the platform take, what do I keep, and what do I have to charge to keep $X?

No dependencies. No build step. The JSON is usable on its own — data/fees.json is the product; index.js is a convenience.

Install

npm install marketplace-fee-data

Or straight from the repo:

npm install github:jelonman/marketplace-fee-data

Or just take the file — the JSON is the whole point and it stands alone:

curl -O https://raw.githubusercontent.com/jelonman/marketplace-fee-data/main/data/fees.json

It is self-describing: data/schema.json is a JSON Schema (draft 2020-12) covering every field, so editors and CI can validate the dataset without this package.

Use

const { feeOn, breakdown, priceForNet, getPlatform } = require("marketplace-fee-data");

feeOn("etsy", 30);
// 3.3   — 9.5% of $30 plus the $0.45 fixed

breakdown("depop", { salePrice: 40, shippingCharged: 5, cost: 12 });
// { gross: 45, fee: 1.94, costBasis: 12, net: 31.06, marginPct: 69.02 }

priceForNet("poshmark", 18, 25);
// 53.75 — list at this and you clear $25 after Poshmark's 20% and your $18 cost

getPlatform("mercari").notes;
// "Mercari reintroduced a seller fee in January 2025: a flat 10% of the item price..."

A platform that is not in the dataset still works — pass the model directly:

feeOn({ percent: 8.5, fixed: 0.35 }, 120); // 10.55

The data

| Platform | id | Headline rate | What that is | Calculator | | --- | --- | --- | --- | --- | | Vinted | vinted | 0.0% | Vinted charges sellers $0 — buyers pay the protection fee | calculator | | Depop | depop | 3.3% + $0.45 | 3.3% payment processing + $0.45 fixed per sale | calculator | | Whatnot | whatnot | 10.9% + $0.30 | 8.0% seller fee + 2.9% payment processing + $0.30 fixed | calculator | | Patreon | patreon | 12.9% + $0.30 | 10% platform fee (new creators, Aug 2025+) + 2.9% + $0.30 | calculator | | Gumroad | gumroad | 12.9% + $0.30 | 10.0% seller fee + 2.9% payment processing + $0.30 fixed | calculator | | DoorDash | doordash | 0.0% | No cut taken from what DoorDash pays you — your costs are gas/vehicle | calculator | | Etsy | etsy | 9.5% + $0.45 | 6.5% transaction + 3% + $0.25 processing + $0.20 listing fee | calculator | | eBay | ebay | 13.6% + $0.40 | 13.6% final value fee + $0.40 per order | | | Poshmark | poshmark | 20.0% | 20% on sales of $15+ — a flat $2.95 below that | | | Mercari | mercari | 10.0% | Flat 10% of the item price — no separate processing fee | |

Percentages apply to the gross (item plus any shipping you charge the buyer), which is how every one of these platforms actually bills. Each entry carries a notes field with the caveats that break naive math — Poshmark's flat $2.95 under $15, eBay's category variance and the 2.35% band above $7,500, Patreon's legacy plans, Etsy's Offsite Ads.

Why this exists

Every "what will I clear on this?" tool re-hardcodes the same numbers, and they rot. Depop dropped its 10% seller fee in 2024. Mercari brought one back in January 2025. Patreon changed platform fees for creators onboarded from August 2025. Nobody is maintaining a shared, machine-readable copy, so everybody maintains a private, stale one.

This is that shared copy. Each row links to the platform's own fee page so you can check it rather than trust it.

Accuracy and scope

These are the common default seller cases in USD. Category, country, subscription tier and legacy plans all change them, and platforms change them without notice. Check the linked source before you price anything on these numbers. If a rate here is wrong, open an issue with a link to the platform's fee page and it gets fixed — that is the entire point of the repo.

Tests

npm test

Twelve tests, no framework, no dependencies. The interesting one is the round-trip: for every platform in the dataset it computes priceForNet, then runs that price back through breakdown and asserts the seller really does clear the target. That catches an inverted fee formula, which a "fee > 0" assertion never will.

The suite also validates data/fees.json against data/schema.json, and then mutates the dataset to prove each rule actually rejects bad input — a validator that never fails is worse than none.

Licence

MIT. The fee figures are facts about public pricing pages; use them however you like.