Somagraph

$SOMAGRAPH is live on Solana. Decode your biological age. 60 seconds. Price of a coffee. Built on peer-reviewed PhenoAge science. Every analysis burns 1,000 $SOMAGRAPH permanently. Every USDC payment triggers buyback-and-burn. The protocol is deflationary from transaction one. somagraph.bio Ayh1gFcpNWhra7Zt1D1aqPF8syk9usaELdAWMEutpump pump.fun/coin/Ayh1gFcpN…

Move sliders for each biomarker. Watch your Bio Age update in real time. No upload. No signup. No wallet required. 100% client-side computation. PhenoAge formula is open math. Zero API cost per interaction. Average time-on-page: 3-5 minutes. Built-in viral hook: screenshot your Bio Age, share to X. $SOMAGRAPH

Six fictional patients with realistic biomarker panels. Pre-cached AI analyses. Zero per-view inference cost. User clicks, instantly sees what a $SOMAGRAPH result looks like. Alex, 32, tech founder, bio age 38. Sarah, 45, marathon runner, bio age 39. Kenji, 28, biohacker, bio age 24. Marcus, 58, diabetic, bio age 67. Different demographics. Self-selection. Emotional engagement before they upload anything. somagraph.bio

The $SOMAGRAPH Anchor program on Solana provides four instructions: initialize_protocol — one-time treasury and config PDA setup. record_analysis — immutable attestation linking wallet, panel hash, PhenoAge, score, timestamp. burn_payment — SPL Token-2022 burn of 1,000 tokens. usdc_buyback_burn — cron-triggered Jupiter swap and burn. Attestation PDAs derived from [b"attestation", wallet, nonce]. Every user's analysis history is enumerable on-chain without exposing any health data. Open-source from day one: github.com/somagraph

At 100 analyses per day with 80/20 token-USDC split: Direct burn: 80,000 $SOMAGRAPH daily. Buyback burn: ~50,000 $SOMAGRAPH daily. Annual: ~47.5M tokens burned. 4.75% of total supply. At 1,000 analyses per day: ~475M tokens burned annually. 47.5% of supply. The token is structurally deflationary when the product has usage. That is the entire thesis. $SOMAGRAPH

Two independent supply-reduction streams power $SOMAGRAPH. Vector 1: User pays 1,000 $SOMAGRAPH. Entire amount burned permanently via SPL Token-2022 burn instruction. Vector 2: User pays $5 USDC. Protocol splits 50/50. Half buys back $SOMAGRAPH on Jupiter every 24 hours and burns it. Half funds operations. User always pays whichever is cheaper in USD terms. Token holders win on the way up. USDC protects accessibility when token moons.

> No raw biomarkers are ever stored in plaintext in the $SOMAGRAPH infrastructure. Not in the database. Not on-chain. Not in logs. Panel data is encrypted with the user's wallet public key before transmission. The Edge Gateway decrypts, processes, and re-encrypts in a stateless worker. No persistent storage of plaintext. > PostgreSQL stores AES-256-GCM ciphertext. Only the user's wallet can decrypt. Right to delete: sign a wallet message, all encrypted records are purged. On-chain: only SHA-256 hashes of normalized marker JSON. Sufficient entropy across 50+ floating-point values makes dictionary attacks computationally infeasible. > This is not HIPAA-certified. We are honest about that. US is geofenced in V0. But the architecture is designed to exceed HIPAA technical safeguard requirements from day one. We build the security first and pursue the certification second. somagraph.bio

PDF upload → format detection → image conversion at 300dpi → Vision AI multimodal OCR → structured JSON extraction → canonical normalization → unit conversion → PhenoAge scoring. 50+ biomarkers extracted from any lab format. Quest. Labcorp. Employer checkup. Hospital discharge. Any format. Cost per parse: ~$0.01. Charge per analysis: $5. That is 99.8% gross margin powering $SOMAGRAPH.

1 How $SOMAGRAPH works under the hood. Five containers. Four trust zones. One pipeline. Upload to result in under 60 seconds. Web app (Next.js 15) → Edge Gateway (Cloudflare Workers) → AI Narrator (Vision OCR + text gen) → Scoring Engine (pure TypeScript math) → Solana Program (Anchor). 2 The Edge Gateway handles rate limiting, wallet signature verification, free-trial eligibility, and geofencing. Every request is authenticated via Ed25519 signature. Anti-sybil gate: wallet age 7+ days, email, IP fingerprint. No shortcuts on access control. This is health data. 3 The Scoring Engine is embedded in the Edge Gateway. Pure math. No external API call. PhenoAge + Klemera-Doubal computations run in-process. The AI layer is separate and stateless. It interprets. It does not compute. That separation matters for reproducibility and auditability. 4 The Solana program stores only SHA-256 hashes of normalized biomarker panels. No PII on-chain. No raw values. Ever. Four instructions: initialize_protocol, record_analysis, burn_payment, usdc_buyback_burn. Don't trust, verify: github.com/somagraph

Every feature in V0 is engineered to drive a single emotional reaction in under 60 seconds. "I am biologically older than I thought." That reaction is what makes the user share, buy more analyses, and tell three friends. Specificity over vagueness. Comparison over isolation. Three priorities, not fifty. $SOMAGRAPH

🩺 Meet Dr. Soma. AI longevity physician living inside somagraph.bio/app. Click the ⚕ bubble — ask anything about your bio-markers, sleep, supplements, protocols. Answers in seconds, cites peer-reviewed evidence. Off-topic? Politely declined. Powered by Claude

Tutorial + curl + SDK examples + error codes: #api" target="_blank" rel="nofollow noopener">somagraph.bio/docs#api Need a key? DM @somagraph or t.me/somagraph. 24h turnaround. Self-service key issuance + wallet-bound quotas ship in v0.5. 🧬 $SOMAGRAPH CA: Ayh1gFcpNWhra7Zt1D1aqPF8syk9usaELdAWMEutpump

What people are building: • Telegram / Discord bots — DM a PDF, bot replies with PhenoAge • Wearable apps — combine Oura sleep with our lab decode • Research pipelines — batch cohort analysis with reproducible Levine math • Personal AI agents — longevity context for Claude / GPT-4

You asked for an SDK. It's live. @somagraph/sdk — TypeScript, zero deps, works in Node 18+ and browsers. Lab PDF → real Levine PhenoAge (2018, peer-reviewed) → AI-narrated priorities. Original ETA was v0.5, ~10 days. Shipped today. Tutorial: #api" target="_blank" rel="nofollow noopener">somagraph.bio/docs#api

Calling biohackers, longevity nerds, and Quest/Labcorp regulars: We want to stress-test our extraction across panel formats. If you upload a PDF and the result looks off — DM us a screenshot of the result + the marker that's wrong. We'll patch the regex same-day. Free PhenoAge + you help build the standard. somagraph.bio/app/upload

"I don't have a lab PDF handy" — fair. Two paths: 1. Take the AI consult: somagraph.bio/app/consult 12 questions → real Gemini-narrated bio-age estimate. ~3 minutes. 2. Browse 6 sample profiles: somagraph.bio/app/demo Bryan Johnson, Peter Attia, Andrew Huberman analyzed via public disclosure. See what a finished result looks like. Either way, you'll see the real engine.

v0.3 just landed. ✅ Live $SOMAGRAPH price/MC/volume widget on LP + app ✅ Decode history persists across sessions ✅ Download your bio-age report as PDF Pipeline ships every few hours. Build in public. somagraph.bio/app CA: Ayh1gFcpNWhra7Zt1D1aqPF8syk9usaELdAWMEutpump