Donation wishes that
route to the right place.

Fundraising platforms collect cash.

Gish funds outcomes. The difference matters more than it sounds.

GoFundMe is the most successful fundraising rail of the past fifteen years and has done meaningful good in the world. It is also, structurally, a campaign-shaped product: a story, a goal, a meter, a payout to the campaign owner’s bank account, and a trust gap between the contributor and the eventual recipient. In the cases where the campaign owner is the recipient (most of them) this works fine. In the cases where it isn’t — a family member raising for a sick child, a friend raising for a neighbor in eviction, a community raising for a family hit by a fire — the trust gap matters. The contributor has no way to verify that the dollars they sent actually reached the cost they were raised against.

For the average campaign that gap is fine. For the campaign that needs to scale beyond a contributor’s immediate trust network — the campaigns that go viral, the campaigns aggregating community money for a stranger, the campaigns where the asymmetry of harm if the funds are misused is real — the gap is the limiting factor. Most contributors who don’t personally know the campaign owner give modest amounts and stop early. The structural reason is trust, and trust can’t be solved with copy or photos.

Gish’s answer is an architectural one. Cause campaigns on Gish route to a verified payee — the hospital, the university, the landlord, the 501(c)(3) — not to the campaign owner. The owner’s job is to identify the right payee and document the outcome; the money settles directly to that endpoint. The contributor sees the destination at the moment of contribution. The owner doesn’t touch the dollars. The trust gap collapses to a verification gap, which is a tractable problem.

Form 990 lookup, at the source.

Every nonprofit recipient is verified against the IRS’s public 990 dataset before payout. Not after; not in batch. Before.

The IRS publishes 990 filings for every registered 501(c)(3). The data is machine-readable, structured, and public. Gish queries it as part of the payee onboarding flow. When a campaign nominates a nonprofit as its outcome payee, we resolve the EIN, pull the latest 990, confirm the organization is in good standing, confirm the payee bank account matches a name on the filing, and only then mark the payee verified. Verification typically completes in 24 to 72 hours; the campaign can collect pledges during this window, but the funds remain escrowed until verification clears.

Tax-deductible receipts auto-generate at the moment of contribution and update if the contribution is matched, increased, or refunded. Year-end tax statements aggregate across a contributor’s full year and ship by mid-January for the prior tax year. The receipt template is co-engineered with two large 501(c)(3) finance teams and survives the IRS substantiation requirements at the $250 threshold without additional paperwork from the contributor.

• EIN → 990 lookup at onboarding. Status, name match, address match, principal-officer match.
• Bank-account name match. Payout account must match a name on the filing.
• Auto-generated tax-deductible receipts. At contribution moment, with $250+ substantiation language.
• Year-end aggregated tax statements. Shipped by January 15 for prior year.
• Anonymous contribution option. The 501(c)(3) sees the contribution and amount; the contributor identity stays masked.
• Re-verification annually. 990s update annually; we re-check status and flag campaigns whose payees fall out of good standing.

Escrow released only to verified payees.

For the cases that aren’t 501(c)(3) — a hospital, a university, a landlord — the same verification logic applies. The destination is named; the funds settle there.

Most fundraising on the consumer internet is for individuals, not nonprofits, and most of it has no defensible destination story. A campaign for a sick child often ends up depositing into a parent’s checking account, and the contributor has no way of knowing whether the dollars reached the medical bills or covered groceries that week. The outcome fund model rejects that ambiguity. The destination is named at the campaign-creation moment — the hospital’s patient billing department, the university bursar, the landlord’s property management company — and the payee is verified the same way a 501(c)(3) is verified: EIN or DUNS or business-registration lookup, address match, bank-account name match. The verification queue runs the same 24 to 72 hours.

The contributor sees the destination at the moment of contribution. The campaign page shows the Trust Score and the verified payee name. Funds remain escrowed until the payee is verified; if verification fails, contributions refund. If verification succeeds, funds release to the payee directly, with the contributor seeing the disbursement event. The campaign owner never holds the money.

The number contributors can see.

An honest signal of how verified a campaign is, surfaced at the moment of contribution.

The Trust Score is the headline signal contributors see when they’re deciding whether to give. It’s computed from a published set of inputs: payee verification status, payee history, campaign-owner identity verification, prior-campaign delivery history, document completeness, and the named-payee match strength. The inputs are auditable; the formula is published; the score updates in real time as new evidence arrives. A campaign whose payee fails verification mid-flight drops in score and surfaces a warning to contributors before they complete a contribution.

A high Trust Score isn’t a promise that the campaign owner is a good person or that the situation is real. It’s a promise that the money will reach the destination the campaign named. That’s the contract Gish can actually keep, and the only one we’ll claim.

Give without signing your name.

Contributors can stay private. The 501(c)(3) still receives. The receipt still issues. The math still works.

Anonymous contribution is a first-class option on every cause campaign. The contributor enters their name and email for receipt-delivery purposes; the campaign owner and the public page see only an anonymous handle. The 501(c)(3) recipient sees the contribution amount and the routing information needed for accounting; depending on the recipient’s reporting standards, the contributor identity may or may not be passed through. The contributor controls the visibility, the law decides the floor.

For tax purposes the receipt is real and named — the IRS substantiation requires it — but the public surface does not have to be. This matters in the cases where the contributor would prefer not to publicly signal a gift, where the contributor is in a relationship with the recipient that makes public attribution awkward, or where the contributor simply doesn’t need credit for the gift.

Against the incumbents.

GoFundMe is the most-used. Classy, Givebutter, Donorbox serve nonprofits well. None of them ship outcome-funds.

GoFundMe’s strength is the campaign-shaped product and the social-graph virality of a story. Gish’s strength is the verified-destination model. They’re different shapes for different cases. If you’re raising for someone you know personally and your contributor network is your friends, GoFundMe’s campaign shape may be fine. If you’re raising for someone you don’t know personally and you need to scale beyond your immediate trust network, the verified-destination model is the only architecture that closes the trust gap.

Transparent. Published.

3% platform fee plus processor. That’s the whole structure.

For comparison: GoFundMe’s platform fee on personal fundraisers is currently 0% (funded by optional contributor tips), with the same processor fee on top; Classy and Givebutter run 2–5% platform fees plus processor. Gish’s 3% is in the middle of the band, with the verification + escrow + Trust Score machinery providing the differentiated value the fee covers.

What this doesn’t do.

A useful list of where the architecture stops.

The verified-payee model works for cases where a clean payee exists. It works for hospitals, schools, landlords, utility companies, 501(c)(3)s, and named contractors. It does not work for cases where the recipient legitimately needs cash for unstructured living expenses across an indefinite window — those campaigns are still better served by traditional cash-collection platforms, and we’re honest about it. The fastest disqualifier is the absence of a payee: if you can’t name where the money needs to land, Gish isn’t the right tool for the moment.

International is partial. US-based 501(c)(3)s are first-class; international charities work case-by-case via fiscal-sponsor relationships. International healthcare and education payees work where the institution has a bank account that can receive US-issued ACH or wire and where the institution’s verification data is accessible. The roadmap covers more international expansion; the current surface is honest about its US-centric footprint.

Crypto contributions are out of scope. Tax-deductibility under US law has open questions for crypto contributions, and we’d rather not be the rail that misroutes a contributor through an unclear regulatory state. The position is conservative on purpose.

Why a graph matters here.

Campaign-shaped products start over every time. Gish doesn’t. Trust compounds.

A campaign-shaped product has no memory. Every fundraiser launches into a blank social space, with whatever copy and photos the owner can muster, against whatever borrowed trust the owner can summon from their personal network. When the campaign ends, the artifacts dissolve. The next campaign starts from zero. This works for the easy case — you know the owner, you know the situation, you give. It fails for the harder cases where a campaign needs to scale beyond an immediate trust network or where a community wants to track an owner’s history of delivering on what they raised against.

Gish’s cause campaigns live inside a persistent identity graph. A campaign owner who has successfully run prior outcome-fund campaigns carries that history forward into their next campaign. A verified payee that has cleanly received and reported on prior funds carries its track record into the next funding cycle. A contributor who has given to dozens of campaigns sees the patterns of how their giving has translated into actual outcomes. None of this is invasive social-graph data; it’s strictly the on-platform delivery history that the system already records as part of normal operation.

The compounding effect is the durable advantage. A new owner running their fifth cause campaign has earned a Trust Score that reflects four successful prior verifications, and contributors give to that campaign with information no first-time campaign could ever surface. A new payee that has cleanly disbursed millions of dollars through dozens of campaigns has built operational credibility that no copy can manufacture. The platform gets more useful over time precisely because the campaigns aren’t the unit of accumulation — the trust graph is.

Treasury pools deepen this further. A family or community Treasury that routinely deploys onto cause-fund payees develops a track record of clean disbursement and verification, and the recipient organizations earn a corresponding history. Over years the graph becomes the substrate for a level of confidence in cross-network giving that campaign-shaped fundraising platforms can’t structurally provide.

What we do with the money in transit.

Escrow, SOC2, processor-level controls, dispute handling.

Funds collected during the verification window sit in a Stripe-managed escrow account, not in a Gish corporate operating account. They are FBO (for benefit of) the eventual payee from the moment they are received. If verification fails or a campaign is canceled, refunds disburse from the escrow directly back to contributors’ original payment methods with no intermediate handling. If verification succeeds, funds release to the payee’s verified bank with a standard ACH or wire timeline.

Disputes — chargebacks, contributor refund requests, payee disputes — route through a dedicated trust-and-safety queue with published SLAs. Stripe’s standard chargeback windows apply; refund requests outside the window are still reviewable but no longer auto-routable. Every dispute decision is logged and the rationale is preserved for the contributor and the payee, so the record of why a particular contribution was clawed back or sustained is auditable years later. We do not accept platform-level liability for payee misconduct after disbursement — that is what the verification step is supposed to prevent up-front — but we do accept liability for our own verification failures, and the platform has a published reimbursement policy for that case.