Most people discover ProofPool through the launchpad — a token already live, already trading, with a charity name attached. It looks simple. That's intentional. But the machine behind that token is anything but.
This article is the third in our series. Article #1 made the case for why AI verification exists. Article #2 explained how the 5-dimension scoring model works. This one covers the whole pipeline end-to-end: from the moment an organization is first identified to the moment a verified meme coin goes live on-chain with transparent fee splits.
Five stages. One automated system. Zero compromises on verification.
The Full Pipeline
Here's how an organization goes from unknown to tokenized — and why each stage exists.
ProPublica Nonprofit Explorer, CoinGecko token registry, or GoFundMe campaigns. The discovery engine scans these continuously, applying filters for organizational size, activity signals, mission keywords, and crypto-relevant indicators. No human curates this list — the engine surfaces candidates autonomously, 24/7.
Every surfaced candidate is scored on Transparency, Efficiency, Impact Evidence, Leadership, and Innovation — each weighted to reflect how crypto-native donors actually assess trust. The composite score runs from 0 to 100. Organizations scoring below 70 are flagged as insufficient and excluded from the launch queue. The full scoring methodology is documented in our scoring article.
AI scoring produces a recommendation, not a final decision. Every organization that clears the scoring threshold sits in a pending approval queue. A human reviewer examines the score breakdown, source data, and any anomalies flagged by the system before approving the organization for tokenization. This is the one intentional bottleneck — it exists because a false positive (a bad actor getting a verified token) does more damage than a false negative (a good organization waiting longer).
Once approved, ProofPool's pump.fun integration creates the token automatically. Ticker, name, and metadata are generated from the organization's profile. The token launches on Solana with a bonding curve — the same mechanics as any other pump.fun token, but with verified charity metadata embedded at creation. The organization's wallet address and Impact Score are included in the token metadata, making the charitable link verifiable on-chain from day one.
Every trade on a ProofPool token generates fees that are split automatically, on-chain. The charity receives the largest share, directly. A portion funds liquidity to keep the token healthy. A small platform fee sustains ProofPool's operations. Every distribution is logged on-chain and visible to anyone — no black box, no "trust us."
Stage 1: The Auto-Discovery Engine
The problem with most charity verification platforms is that they rely on self-reporting — organizations apply, and the platform evaluates who shows up. That creates selection bias. The organizations most likely to apply are the ones with marketing teams, not necessarily the ones doing the best work.
ProofPool's auto-discovery engine inverts this. It actively scans three data sources to surface candidates:
ProPublica Nonprofit Explorer provides IRS Form 990 data for over 1.8 million US tax-exempt organizations. The engine filters for revenue above $100K (indicates real operations), program service ratios above 60%, and mission keywords across global health, poverty alleviation, environmental impact, and education. A high-performing medical charity in rural Kenya with no marketing budget is just as likely to surface as a well-funded US-based organization with a PR team.
CoinGecko surfaces the other end of the funnel — tokens already marketed as charity or impact coins. The engine ingests token metadata, cross-references claimed beneficiary organizations, and flags mismatches between on-chain fund flows and marketing claims. Many tokens that call themselves "charity" tokens have never transferred a single dollar to any beneficiary. The engine catches that.
GoFundMe represents the long tail: grassroots projects that may lack formal nonprofit status but demonstrate genuine community support and measurable outcomes. This source disproportionately surfaces crypto-native and emerging-market initiatives — exactly the organizations that established verification platforms overlook.
Stage 2: AI Scoring in Practice
Every auto-discovered organization gets the same evaluation as a self-applicant. There's no fast-track for well-known organizations, no penalty for being unfamiliar.
The AI evaluates five dimensions — Transparency (25%), Impact Evidence (25%), Efficiency (20%), Leadership (15%), Innovation (15%) — and produces a composite 0–100 score. We covered the full methodology in How Impacta AI Scores Organizations, but the key operational point is what the score triggers:
Score ≥ 90: Qualified for the Verified Impact badge. Prioritized in the approval queue. These are the organizations ProofPool is built for.
Score 70–89: Qualifies for review. Good organizations with room for improvement in one or two dimensions. May launch without the Verified Impact badge until scores improve.
Score < 70: Excluded from the launch queue. Flagged for re-evaluation if the organization later improves its transparency or impact documentation. No permanent bans — just current-state assessments.
Scores aren't static. An organization that publishes audited financials triggers a re-evaluation. One that goes dark — stops reporting, removes wallet disclosures — will see its score fall in the next cycle. The system reflects current reality, not historical reputation.
Stage 3: Human Review — the Intentional Bottleneck
ProofPool is deeply automated. But one human decision sits between AI scoring and token launch, and it's there for a reason.
The pending approval dashboard shows every organization that has cleared the scoring threshold. Reviewers see: the full score breakdown by dimension, the source data the AI used, any anomaly flags (unusual fund flows, mismatched public records, recent leadership changes), and the proposed token metadata.
Reviewers can approve, reject, or flag for additional research. Rejections are documented with reasons, which feeds back into the scoring model to improve future evaluations. The goal isn't just to catch bad actors — it's to make the AI smarter with every cycle.
Average time in the approval queue is under 48 hours. For organizations with Verified Impact scores (90+), it's typically same-day.
Stage 4: Token Creation via pump.fun
This is the part that surprises people. ProofPool doesn't run its own token launchpad from scratch — it integrates with pump.fun, Solana's dominant meme coin creation platform. That's an intentional technical and distribution decision.
pump.fun has the deepest liquidity, the most active traders, and the most robust bonding curve mechanics on Solana. Building a parallel launchpad would mean asking crypto traders to use a less liquid, less familiar venue — for no technical reason other than brand separation. ProofPool charity tokens compete for attention in the same feed as every other pump.fun launch, which means they have access to the same organic discovery and trading volume.
What ProofPool adds to the pump.fun base is the on-chain verification layer: the organization's Impacta AI score, their verified wallet address, their ProPublica or GoFundMe source link, and the fee split contract. These are embedded in the token metadata at creation and visible to any trader who looks.
The result: a meme coin that trades like a meme coin, but comes with a verifiable paper trail to real-world impact.
Stage 5: Transparent On-Chain Fee Splits
Every ProofPool token generates trading fees. Here's exactly where those fees go:
Trading Fee Distribution (per transaction)
60% direct to charity. Routed to the verified wallet address embedded in the token metadata. No intermediaries, no treasury, no "committee decides when to distribute." Every fee cycle, 60% moves directly to the beneficiary organization's Solana wallet. They can see it arrive. So can you.
25% to liquidity. Meme coins die when liquidity disappears. Routing a quarter of fees back into the pool creates a flywheel: more volume generates more liquidity, which reduces slippage, which attracts more volume. This is the mechanism that gives ProofPool tokens longevity beyond the initial launch spike. It's also what differentiates them from rugpulls — funds that go back into liquidity can't be extracted by founders.
15% platform fee. This sustains ProofPool's infrastructure — the discovery engine, the AI scoring pipeline, the human review process, and the on-chain contracts. It's the lowest viable take rate for a sustainable operation, and it's disclosed upfront in every token's metadata.
What Organizations Need to Do
Here's the most common question from nonprofits discovering ProofPool for the first time: What do we have to do?
The honest answer is: not much, and sometimes nothing at all.
Auto-discovered organizations don't need to apply. If the discovery engine surfaces them and the AI score clears the threshold, they move through the pipeline. ProofPool reaches out to notify the organization before launch and to confirm their wallet address for fee distribution. But the process doesn't wait on a response — tokens can launch with a verified wallet address the organization confirms later, with fees held in escrow in the interim.
Organizations that want to apply directly can do so at /apply.html. The pipeline is the same — the only difference is that the organization provides source documents directly rather than the engine pulling them from public records. Self-applicants typically move through human review faster because the data is cleaner.
Your organization might already be in the pipeline.
The auto-discovery engine scans continuously. If you run a verified nonprofit or impact project, you may already have a score. Check the launchpad to see what's live — or apply to expedite your evaluation.
Why This Pipeline Matters
Every component of this pipeline exists to solve a specific failure mode in existing crypto charity.
The auto-discovery engine solves selection bias — it catches the organizations that wouldn't have applied. AI scoring solves inconsistency — every organization gets evaluated on the same criteria with the same rigor. Human review solves edge cases — the AI catches 95% of bad actors; humans catch the rest. pump.fun integration solves distribution — verified tokens compete in the same venue as everything else, not a walled garden. Transparent fee splits solve accountability — on-chain distribution means donors can verify their impact without trusting a dashboard.
None of these problems is novel. We've known for years that crypto charity has a trust deficit. The pipeline isn't a vision statement. It's the operational answer to that deficit, built to run without human intervention at every stage where human intervention creates bottlenecks or corruption risk.
Article #1 was the why. Article #2 was the how we score. This article is the how it all works.
Now you know the machine. The launchpad is the output.