Documentation

Eluent takes a single-letter amino acid sequence and returns:

• Identity — what the peptide is, with a confidence grade and citations.
• Structure — a 3D model from AlphaFold, ESMFold, or PEP-FOLD4 routed by sequence length.
• Mechanism — primary targets, mechanism summary, links to GtoPdb / ChEMBL / OpenTargets.
• Vendor matches — when our nightly scrapers find this exact sequence on a Tier-1 vendor with a recent COA, the row appears in your dossier.
• On-chain receipt — every lookup is anchored to Solana mainnet via the Memo program with a commit/reveal scheme. Verifiable by anyone, post-hoc.

The free tier uses a single Eluent paymaster account to sponsor the chain writes. You never need a wallet to use the lookup — connecting one only unlocks tier-gated features.

Lookup runs as five sequential stages. Cache hits short-circuit later stages. SHA-256 of the canonicalized sequence is the cache key — same sequence always returns the same dossier.

1. 01CanonicalizeParse FASTA / raw / URL · normalize · compute SHA-256 · check cache
2. 02IdentityUniProt Peptide Search exact (≥7aa) → BLAST homolog (e-val ≤ 1e-10) → grade
3. 03StructureAlphaFold DB v6 / ESMFold / PEP-FOLD4 routed by length + UniProt match
4. 04MechanismChEMBL + GtoPdb + OpenTargets join · LLM synthesis · cite-or-omit
5. 05On-chainCommit memo before bio APIs · reveal memo with result CID after upload

Every dossier carries one of five letter grades — A, B, C, D, or X. Same scale on the result page, the OG share image, and the embed badge. The grade encodes how much the sequence is trusted, not whether the peptide is good or bad.

Full rubric: /docs/grading.

Each lookup produces two transactions on Solana mainnet-beta — a commit before the bio APIs run, then a reveal after the dossier is uploaded to permanent storage. The commit is a blinded hash; the reveal carries the salt and the result CID. Anyone can later verify the binding without needing Eluent's cooperation.

// COMMIT memo (lands at confirmed, ~400ms)
eluent:commit:v1:<sha256(sequence||salt||requested_at)>:<requested_at>

// REVEAL memo (lands after dossier upload)
eluent:reveal:v1:<commitTxSig>:<salt_hex>:<result_cid>:<result_hash>

Cost: ~$0.0015 per lookup (commit + reveal Memo writes). Free-tier writes are sponsored by the Eluent paymaster account; connected-wallet writes can be self-paid for users who prefer full self-custody.

Eluent uses Sign-In with Solana (SIWS) — connect Phantom, Solflare, or Backpack via the Wallet Standard, sign one nonce, receive a JWT scoped to your pubkey. No password. No email.

The wallet button is in the top-right nav. Free anonymous use (no wallet) gives 5 lookups per IP per 30 days; signing in lifts that to a higher per-wallet cap and unlocks the saved-receipts history at /me.

Pro features unlock based on your $ELUENT balance — the contract address ships when the token launches. Until then, all wallets resolve to the free tier.

Drop your raw 23andMe / AncestryDNA / MyHeritage / LivingDNA / FamilyTreeDNA export. Eluent matches your genotype against a curated PGx variant table for the top-12 research peptides. Results render entirely in your browser.

Defense layers stacked behind that promise:

• No upload endpoint exists for genome data — there is no/dna-match/upload route.
• Parsing happens in a Web Worker isolated from the page.
• A route-scoped service worker actively rejects POSTs from /dna-match as a second line of defense.
• Optional cache lives in IndexedDB with AES-256-GCM encryption, deleted on demand from the page.
• Share OG images are CI-checked to contain zero rsID patterns — the share preview cannot leak genotype even by accident.

Open it at /dna-match.

Add multiple peptides to evaluate interactions across the stack. The engine uses three signals — shared receptor competition, shared metabolic pathway, and curated literature stacks — to surface synergies, warnings, and unknowns.

Default posture on unmatched combinations is no known interaction data — Eluent never implies a stack is safe, only what is and isn't in our literature.

Open it at /stack.

• No genome upload. DNA Match parsing is browser-only. There is no server endpoint for raw genotype.
• Sequence privacy. Lookup sequences are committed via blinded hash before any third-party API runs — UniProt sees the sequence on reveal-only paths or not at all (cache hits skip the API entirely).
• No medical PII. Eluent stores no name, email, address, or birthdate. Wallet pubkey is the only identifier.
• Analytics whitelist. Captured events are restricted to a fixed allowlist. rsIDs, genotype calls, and evidence-level details are explicitly denied at the capture layer.
• No Eluent-side keys. Wallet signing happens in the user's wallet extension. Eluent never sees a private key.
• Verifiable on-chain. Every lookup binds the sequence to a Solana mainnet transaction via commit/reveal — anyone can later verify the dossier against the on-chain hash without trusting Eluent's servers.

• Research only. Not for human use. Persistent disclaimer in the footer of every page that lists a peptide.
• No medical claims. Mechanism descriptions cite primary literature; treatment recommendations are out of scope.
• No vendor rankings. The vendor table shows price + COA freshness — never a "best of" or "buy now" rail.
• No checkout, no broker, no affiliate links. Eluent is a pure information layer.
• Vendor scrapers identify honestly (Eluent-VendorBot/1.0), respect robots.txt, throttle, and never republish vendor content verbatim — only price + outbound COA link.
• 18+. DMCA agent registered. 21 CFR aware.