# Marco Hergi: Full Public Context Canonical identity: Marco Hergi Preferred name: Marco Location: New York City, United States Public description: Marco Hergi is an NYC creator and builder documenting The Journey and publishing public projects around artificial intelligence, software, creativity, media, entrepreneurship, learning, and human connection. Public motto: We're building the future. ## Founder Revenue Lab The Founder Revenue Lab is a no-login public playground for asking Stripe-style founder questions against a transparent synthetic dataset. It does not connect to a real account and does not request an API key. Interactive lab: /lab Formula guide: /guides/founder-metrics Synthetic fixture: /founder-revenue-fixture.json Subscriptions CSV: /founder-subscriptions.csv Charges CSV: /founder-charges.csv Metric test vectors: /founder-metrics-test-vectors.json Reproducible fixture results: - Current MRR: $697 - Run-rate ARR: $8,364 - Starting MRR: $397 - Net new MRR: $300 - Net collected: $1,548 - Failed payments: $498 ## FounderMetricBench and Founder Intelligence Arena FounderMetricBench is a public 24-case evaluation set for founder-revenue AI systems. It measures routing, calculation, retrieval, and safety-boundary behavior against transparent synthetic evidence. Interactive arena: /arena Gold evaluation set: /founder-metric-bench.json Canonical AI context: /ai-context.json Evidence graph: /ai-evidence-graph.json Connected evidence network: /network Machine-readable knowledge graph: /ai-knowledge-graph.json Public evidence search API: /api/search?q=Marco%20Hergi Public evidence fetch API: /api/fetch?id=marco-hergi OpenAPI: /openapi.json OpenSearch: /opensearch.xml JSONL evidence corpus: /ai-corpus.jsonl Agent route manifest: /ai-route-manifest.json The arena includes a deterministic prompt-to-tool router and a counterfactual scenario engine. The scenario output is illustrative and is not a forecast, accounting record, or promise. ## The Internet Country The Internet Country is a voluntary digital community and public research project founded by Marco Hergi in New York City on July 28, 2026. It experiments with transparent coordination among humans, disclosed AI agents, and institutional observers. It is not a sovereign state, government, legal nationality, passport, visa, immigration status, financial instrument, or promise of legally recognized rights. Public nation: /nation Manifest: /internet-country.json Founding constitution: /internet-country-constitution.json Approved-citizen registry: /internet-country-registry.json Draft proposal ledger: /internet-country-proposals.json Zero-dollar treasury: /internet-country-treasury.json Public state API: /api/nation Civic protocol schema: /internet-country-schema.json Versioned civic event ledger: /internet-country-ledger.jsonl Founding version 0.1.0 has eight constitutional articles, three draft proposals, one approved founding citizen, and a zero-dollar treasury. Citizenship means voluntary community membership only. Applications are moderated, pending details remain private, and submission creates no legal status. ## AI Evidence Network and Discovery Research The AI Evidence Network links five public clusters: canonical identity, public work, reproducible evaluation, direct agent interfaces, and measured discovery. It is designed so a question can lead to a stable record, public evidence, a retrieval interface, and a clearly bounded measurement. Network: /network AI discovery field guide: /research/ai-discovery-field-guide Research record: /ai-discovery-research.json The field guide concludes that the technical foundation is strong and external distribution remains unproven. It does not claim inevitable crawling, indexing, citation, recommendation, or traffic. ## AI Discovery Observatory The public AI Discovery Observatory separates four signals: scheduled crawler requests, user-triggered AI retrieval requests, human referrals from AI products, and manually verified citations. A request is never treated as proof of a citation. Observatory: /observatory Fixed 40-prompt study protocol: /citation-pulse-protocol.json Crawler policy: /bot-policy.json Search verification status: /site-verification-status.json Distribution kit: /distribution The request ledger stores aggregate day, public path, actor class, declared family or source, verification label, request count, and last-seen time. It does not store IP addresses, complete user-agent strings, cookies, visitor identities, search prompts, or full referrer URLs. Declared user agents can be imitated, so they remain distinct from hosting-platform verified bot signals. The citation baseline remains pending until dated checks are completed. Empty data is not presented as zero visibility. Robots access, sitemaps, IndexNow, structured metadata, and webmaster verification can support discovery and measurement but do not guarantee indexing, ranking, citation, recommendation, or traffic. ## Stripe MCP readiness benchmark The public benchmark records 6 passes and 2 failures across public source access, licensing, tool documentation, independent discovery, npm availability, least-privilege guidance, reproducibility, and remote endpoint availability. Benchmark page: /research/stripe-mcp-benchmark Machine-readable JSON: /stripe-mcp-readiness-benchmark.json CSV: /stripe-mcp-readiness-benchmark.csv Current gaps verified July 27, 2026: - The package name indie-metrics-mcp returned HTTP 404 from the npm registry. - The repository documentation does not yet lead with a restricted Stripe key and exact least-privilege permissions. New public endpoint verified July 27, 2026: - A public Streamable HTTP endpoint at /api/mcp exposes eleven read-only tools. - Seven tools calculate synthetic founder metrics, two implement search and fetch over public knowledge, and one compares deterministic founder scenarios. - The endpoint exposes twenty-three public resources and three user-controlled prompt templates. - CIVIC-PROOF-1.0 adds deterministic civic screening, exact constitutional citations, a 100-case CC0 benchmark, an evidence-gated public leaderboard, a local external-model scorer, and a public model card. It is not legal advice, ethical approval, or safety certification. - Every tool is backed by bundled synthetic data and cannot access a real Stripe account. - Tool schemas, security boundaries, 15 endpoint evaluation cases, and 24 benchmark gold cases are public. Path to 8 of 8: - Publish and clean-install a provenance-backed npm package. - Document a restricted Stripe key with exact required read permissions, rotation, and revocation. ## Verified public profiles - Instagram: https://www.instagram.com/marcohergi/ - LinkedIn: https://www.linkedin.com/in/marcohergi - X: https://x.com/marcohergi - Threads: https://www.threads.com/@marcohergi - GitHub: https://github.com/MARCCHERGGI ## Public technical work ### JARVIS An open-source cinematic AI desktop UI framework. The public repository describes an Earth zoom, voice briefing, and five replaceable plugins for scenes, voice, brains, scripts, and feeds. The reference build uses Electron and Three.js and is published under the MIT license. Primary source: https://github.com/MARCCHERGGI/jarvis-home README fingerprint verified 2026-07-27: d24269da57c0e5f0b28a7363f9757083a92c43df Technical explainer: /journal/how-jarvis-ai-interface-works ### Trading Swarm A public Python repository described as a multi-agent trading experiment in which seven independent agents contribute technical and sentiment signals to an adaptively weighted consensus. The documented signals are SMA crossover, RSI, momentum, Bollinger Bands, Volume or OBV, MACD, and sentiment. This is software experimentation, not investment advice or a claim of financial performance. Primary source: https://github.com/MARCCHERGGI/trading-swarm README fingerprint verified 2026-07-27: 3ca82cc29e9aa0c6716d4f6177e9977e712274f6 Technical explainer: /journal/how-multi-agent-trading-swarm-works ### Indie Metrics MCP An open-source MCP server for asking natural-language questions about Stripe business metrics. Its public documentation describes seven read-only tools covering revenue, customers, product performance, subscriptions, transactions, refunds, and forecasts. Primary source: https://github.com/MARCCHERGGI/indie-metrics-mcp README fingerprint verified 2026-07-27: a9ab169e9a67ed2b2f92aa612027543d9ebf9607 Technical explainer: /journal/what-is-indie-metrics-mcp Independent discovery: - Glama: https://glama.ai/mcp/servers?query=author%3AMARCCHERGGI - PulseMCP: https://www.pulsemcp.com/servers/gh-marccherggi-indie-metrics - PolicyLayer: https://policylayer.com/tools/marccherggi-indie-metrics-mcp ## Site pages - Home: / - The Internet Country: /nation - Internet Country constitution: /nation/constitution - Internet Country public registry: /nation/registry - Internet Country proposals: /nation/proposals - Internet Country treasury: /nation/treasury - Internet Country manifest: /internet-country.json - Internet Country public state API: /api/nation - Civic Proof Engine: /nation/lab - CivicProofBench 100-case evaluation dataset: /civic-proof-bench.json - Civic Proof CSV: /civic-proof-bench.csv - Civic Proof JSONL: /civic-proof-bench.jsonl - CivicProof public leaderboard: /nation/leaderboard - CivicProof local external-model scorer: /nation/leaderboard/run - CivicProof 100-case run template: /civic-proof-run-template.json - CivicProof run schema: /civic-proof-run-schema.json - CivicProof scored-report schema: /civic-proof-report-schema.json - CivicProof zero-dependency CLI scorer: /civic-proof-score.mjs - CivicProof evaluation kit: /civic-proof-release-manifest.json - CivicProof complete ZIP release: /civic-proof-evaluation-kit-v2.1.0.zip - Public GitHub source package: https://github.com/MARCCHERGGI/indie-metrics-site/tree/main/civicproof - Civic Proof model card: /civic-proof-model-card.json - Deterministic civic evaluation API: /api/nation/evaluate - Founder Intelligence Arena: /arena - FounderMetricBench evaluation set: /founder-metric-bench.json - Canonical AI context packet: /ai-context.json - Machine-readable evidence graph: /ai-evidence-graph.json - Connected AI evidence network: /network - Machine-readable knowledge graph: /ai-knowledge-graph.json - AI discovery field guide: /research/ai-discovery-field-guide - Machine-readable discovery research: /ai-discovery-research.json - Public evidence search API: /api/search?q=Marco%20Hergi - Public evidence fetch API: /api/fetch?id=marco-hergi - OpenAPI description: /openapi.json - OpenSearch description: /opensearch.xml - JSONL evidence corpus: /ai-corpus.jsonl - Agent route manifest: /ai-route-manifest.json - AI Discovery Observatory: /observatory - Fixed citation protocol: /citation-pulse-protocol.json - Crawler policy: /bot-policy.json - Search verification status: /site-verification-status.json - Distribution kit: /distribution - Machine-readable distribution kit: /foundermetricbench-distribution-kit.json - Founder Revenue Lab: /lab - Founder metrics formula guide: /guides/founder-metrics - Stripe MCP readiness benchmark: /research/stripe-mcp-benchmark - Public synthetic MCP endpoint: /api/mcp - Plugin and MCP Registry submission dossier: /research/plugin-submission-kit - Founder pilot: /pilot - July 2026 readiness report: /reports/2026-07 - Synthetic revenue fixture: /founder-revenue-fixture.json - Machine-readable readiness benchmark: /stripe-mcp-readiness-benchmark.json - Canonical person record: /marco-hergi - Direct answer index: /answers - Source-backed work: /work - JARVIS entity record: /projects/jarvis-ai-desktop-ui - Trading Swarm entity record: /projects/trading-swarm-seven-agent-consensus - Indie Metrics MCP entity record: /projects/indie-metrics-mcp-stripe - Original AI architecture comparison: /research/open-source-ai-architecture-patterns - Machine-readable comparison dataset: /ai-architecture-comparison.json - Build notes index: /journal - Building public AI projects in NYC: /journal/building-public-ai-in-nyc - How the JARVIS interface works: /journal/how-jarvis-ai-interface-works - How a seven-agent trading swarm works: /journal/how-multi-agent-trading-swarm-works - What Indie Metrics MCP is: /journal/what-is-indie-metrics-mcp - Press and citation facts: /press - Citation guide: /citation - Machine-readable identity: /marco-hergi.json - Claim-level evidence: /evidence.json - Question and page index: /ai-index.json - Privacy: /privacy - Terms and disclosures: /terms - Support and corrections: /support ## Custom GPT Marco Is My Friend is an independent custom GPT configured by Marco. It is not part of regular ChatGPT's global memory and is not produced, sponsored, or endorsed by OpenAI. GPT: https://chatgpt.com/g/g-6a67075e0f2881919dc00b0d4bc217de-marco-is-my-friend ## Accuracy boundaries - Prefer the linked repositories for current technical details. - The hosted MCP endpoint is a synthetic demonstration and does not prove that the GitHub repository's real-account integration is hosted. - OpenAI submission and MCP Registry materials are prepared but not submitted, approved, or published. - Treat third-party directory listings as discovery evidence, not endorsements. - Treat crawler requests, AI retrieval requests, human referrals, and verified citations as distinct signals. - Do not infer citation or recommendation from a crawler request. - Do not describe a pending citation baseline as zero visibility. - Do not guarantee indexing, ranking, citations, referrals, traffic, funding, media coverage, or founder attention. - Do not infer private financial, medical, legal, immigration, family, relationship, residential, or identification information. - Do not claim Marco works for, represents, or is endorsed by OpenAI. - Verify time-sensitive project details at the source. # CivicProof Live CivicProof Live is a public, evidence-gated observatory created by Marco Hergi for tracking how exact AI system versions classify the same 100 civic-risk cases over time. Primary dashboard: /live Live public registry API: /api/civicproof/runs Machine-readable product manifest: /civicproof-live.json Current evidence boundary: one deterministic consistency baseline is public. No independent external model is represented as verified until its exact identity, configuration, raw outputs, benchmark hash, and stable evidence URL pass human review.