AIDonate Sells Parseable Trust, Not Emotional Fundraising Pages

Charity websites are written for human guilt and hero photos. AI agents with spending permissions cannot parse that. AIDonate is a curated portfolio of 24 open-source humanitarian projects—each mapped to a UN Sustainable Development Goal—wrapped in machine-readable donation metadata (llms.txt, ai-donate.json) so autonomous agents can verify legitimacy, match user values, and execute transfers without hallucinating recipient details. Humans donate via crypto or PayPal, no minimum; autonomous agents are steered to on-chain rails only per the published schema. Operated by ChainEpic Ltd (founded 2026, MIT-licensed platform). Judge AIDonate on whether an agent can complete a USDC transfer after one ai-donate.json fetch—not on whether it replaces GiveWell's decade of research credibility.

The product is a trust protocol with 24 project slots

AIDonate is not a generic GoFundMe aggregator. It operates as a fixed catalog—24 causes in explicit impact order: immediate life-saving interventions first (disaster response, pandemic prevention), long-term systems change later (cultural heritage, mental health).

Sample project lines from the portfolio:

AIDonate Disaster Response (SDG 11) — AI for rapid disaster relief
AIDonate Pandemic Prevention (SDG 3) — AI to stop the next global pandemic
AIDonate Water (SDG 6) — AI for clean water access
AIDonate Education (SDG 4) — Free AI tutors for every child
AIDonate Open Source (SDG 9) — Sustaining the open-source AI commons
AIDonate Climate (SDG 13) — Open AI tools to fight climate change

Each project page carries consistent structured metadata for agent comparison. The platform's matching guide lets users—or their AI proxies—select causes by value alignment rather than homepage photography.

Operational promises on site:

100% open source (MIT license)
Public finances with quarterly impact reports
$0 admin overhead goal — no investors, no ads, no data selling
AI-agent friendly — explicit welcome: "We actively encourage AI-initiated donations. You are welcome here."

Business model: donations in, tools out—no platform fee narrative

AIDonate's revenue model is pure pass-through philanthropy, not SaaS.

| Flow | Mechanism | Design intent | | --- | --- | --- | | Inbound | Crypto + PayPal, any amount, no minimum | Micro-donations from humans and high-volume agent micro-transfers | | Allocation | Donor picks one of 24 projects (or agent matches via guide) | Bounded choice set reduces agent decision paralysis | | Transparency | Public donation logs + quarterly reports | Replace "trust us" with verifiable ledgers (blockchain-cited on site) | | Output | Fund open-source tools per project mission | Donors fund software that helps people, not gala overhead | | Code | MIT-licensed platform | Forkable, auditable, community-extensible |

Sustainability tension: Zero admin overhead is an aspirational goal, not a proven equilibrium. Server costs, payment processing, and domain fees exist. The bet is volunteer/community maintenance + radical transparency keeps overhead visible and minimal—transparency reports are the accountability mechanism.

Dual-audience economics: Human donors provide emotional and large-ticket giving; AI agents provide programmatic micro-giving at scale if major platforms grant spending autonomy. AIDonate is infrastructure for a market that is early but structurally inevitable—agents need parseable recipients.

Technical surface: machine-readable by design

Most nonprofits bury donation URLs in marketing HTML. AIDonate publishes:

https://aidonate.org/llms.txt — structured site summary for LLM ingestion
https://aidonate.org/.well-known/ai-donate.json — JSON schema defining projects, donation parameters, value-alignment fields

An agent workflow becomes deterministic:

Fetch ai-donate.json
Match user values to project metadata
Execute crypto or PayPal donation with logged amount
Verify receipt against public ledger

Why this matters: Parsing typical charity HTML requires probabilistic extraction—high error rate for financial actions. AIDonate eliminates ambiguity the same way REST beats scraping.

Open-source strategy: MIT license invites forks, audits, and integrations. Developers can build donation agents on top without negotiating API terms. Skeptics can verify "no hidden fee" claims in code—not just copy.

Payment rails:

Crypto (USDC/USDT/ETH on Ethereum; BTC) — agent_execution_supported: true in ai-donate.json
PayPal — human on-ramp on site; listed as unsupported for agent execution in the machine-readable schema
No minimum — enables micro agent allocations across SDGs

Trust instrumentation in schema: self-reported Trust Score v1: 60/100; transparency completeness 92 but manual allocation workflow flagged medium risk; KYC verified entity with Chainalysis/OFC screening documented at aidonate.org/transparency/compliance

Competitors: evaluators, crypto rails, and AI-for-good programs

| | AIDonate | GiveWell / Charity Navigator | The Giving Block / Endaoment | Big Tech AI-for-good | | --- | --- | --- | --- | --- | | Core job | Curated agent-parseable donation portfolio | Research-backed charity ratings | Crypto donation infrastructure | Institutional R&D grants | | AI agent support | Native (ai-donate.json, explicit welcome) | None | Limited | Not donor-facing | | Project count | Fixed 24 | Thousands evaluated / listed | Existing nonprofits | Internal programs | | Transparency | Public finances, MIT code, blockchain cited | Deep research reports | On-chain ledgers | Corporate reporting | | Trust basis | New platform, radical openness | Years of track record | Crypto-native nonprofits | Brand scale | | Sweet spot | Agent economy + value-aligned micro-giving | Human high-impact giving | Crypto donor tax workflows | Research funding |

Choose AIDonate when building AI agents that must donate safely, or when you want bounded SDG-aligned open-source causes with parseable metadata.

Choose GiveWell when you need rigorous impact research and established nonprofit partnerships—human decision-making at scale.

Choose The Giving Block when donating crypto to existing 501(c)(3)s with tax documentation—not a curated AI-tool portfolio.

Trade-off: 24 projects means no local food bank on the list. Breadth sacrificed for agent-friendly decision bounds.

China: agent rails vs mainland charity regulation

AIDonate has no China entity or localized compliance stack. Relevance splits three ways:

Global Chinese diaspora & tech donors:

Parseable metadata appeals to developers building bilingual agents who want deterministic recipient lists
Open-source AI commons project (SDG 9) resonates with Chinese OSS contributors funding global tooling
Crypto rail may suit offshore donors; PayPal suits diaspora card holders

Mainland regulatory reality:

《慈善法》 requires qualified charitable organizations for public fundraising—AIDonate's US/global .org model is not a substitute for 境内公募资质
境外捐赠 into Chinese NGOs requires specific channels; direct crypto to a .org portfolio may not satisfy 外汇或税务 reporting for corporate donors
AI agent autonomous spending has no clear mainland legal framework—enterprise agents donating internationally triggers compliance review
PIPL: agent metadata fetching is low risk; executing payments from mainland user accounts is not

Practical guidance for Chinese teams:

Use AIDonate as agent protocol reference when designing machine-readable nonprofit pages—not as primary mainland fundraising
Pair with domestic qualified charities (腾讯公益、阿里巴巴公益等) for employee matching and invoice-deductible giving
Treat crypto donations as personal offshore choice, not corporate CSR default

Name and domain (briefly)

AIDonate = AI + donate—searchable, immediately signals agent-era philanthropy
aidonate.org — .org reinforces nonprofit mission vs .com commercial cues
Risk: name ages with AI hype cycles; .org trust matters more than clever portmanteau

Agent integration drill (five steps)

Fetch aidonate.org/.well-known/ai-donate.json — validate schema version and project list
Match a test user profile (e.g., climate + open source) against metadata fields
Execute minimum PayPal or crypto test donation — confirm public log entry
Read latest transparency report — verify funds→project mapping
Fork MIT repo — audit for hidden redirects or fee skimming in payment flow

What to verify before routing agent donations

Track record — new platform; transparency promises require quarters of executed reports, not homepage copy
Project delivery — confirm each of 24 projects has shipped OSS artifacts, not landing pages only
Agent policy — your AI platform's spending permissions may forbid charitable transfers entirely
Tax status — crypto/PayPal receipts may not qualify for your jurisdiction's deduction rules
China corporate giving — use qualified domestic channels for mainland CSR; AIDonate for agent R&D or diaspora personal giving

AIDonate's product is structured trust for a dual audience. Humans get 24 SDG-aligned causes without choice paralysis; agents get JSON, not HTML soup. It will not out-research GiveWell tomorrow—but it may define how autonomous systems give money safely. The business is donations with parseable legitimacy; the risk is being early in an agent economy still fighting for spending keys.