Pure Rawz
- Cadence
- Daily
- Scope
- Full catalog
- Last successful refresh
- Notes
- Runs at 03:00 UTC.
How our rankings work
We separate price, stock, testing, transparency, and affiliate relationships so users can evaluate trade-offs.
On this page
Ranking summary
- Collect public vendor offers (product title, size, price, currency, stock).
- Normalize each offer to a price-per-milligram so vial sizes compare cleanly.
- Bucket by in-stock status — in-stock listings sort above out-of-stock.
- Within each bucket, sort ascending by normalized price-per-mg.
- Surface vendor trust signals (COA, third-party tests, transparency status) alongside price.
- Re-run the pipeline on a per-vendor cadence so listings stay current.
Default sort
On peptide pages and best-value comparisons, the default sort is in-stock listings first, then normalized price per milligram, lowest first. Vial sizes vary across vendors; price-per-mg is the only fair comparison.
See an example on the comparison hub.
Transparency scoring (beta)
Composite transparency scoring is in beta. Until it ships, vendor pages display a transparency status drawn from individual signals rather than a single composite number.
Each vendor surfaces one of the following statuses based on the signals below:
- Complete — most signals on file.
- Partial — some signals on file.
- Pending — not yet scored.
- Insufficient data — not enough public information to evaluate.
Signals we use, drawn from each vendor's public site:
- Whether a Certificate of Analysis (COA) is publicly linked
- Whether the COA is lot-specific or general
- Whether a third-party lab is named
- Whether the report date is visible
- Whether purity and identity data are visible
- Product detail completeness
- Refund and shipping policy clarity
- Vendor consistency over time
Third-party verification
On best-value comparisons we show a per-row Third Party Verifiedcolumn. A row marked "Verified" means an independent third-party test record exists for this vendor-peptide pair, aggregated from Finnrick Analytics. It does not guarantee the current batch.
What qualifies: an independent lab report tied to a specific vendor-peptide pair, with a visible report date and identifiable lab. What does not qualify: vendor-published test images without an independent lab named, marketing claims of testing without a report, or self-issued COAs.
Historical tests may not reflect the inventory shipping today. "No test on file" means we do not have a record — not that the product is untested or impure.
Affiliate policy
Affiliate relationships do not affect rankings. Commission rates, payout terms, and whether we have an affiliate relationship at all are not inputs to the sort order or the transparency status.
Sponsored placements, when present, must be labeled and ranked separately. There are no sponsored placements on the site today.
The current affiliate-relationship list is published on the affiliate disclosure page. We continue to track and list vendors without affiliate relationships when they are part of our coverage set.
Data freshness
Vendor data is refreshed on a vendor-specific schedule. Some vendors are checked daily; others weekly, depending on site structure, robots.txt, and crawl limits. Each listing shows its latest verification timestamp when available. Stale offers are archived after 30 days without verification.
Source tiers
Quantitative inputs across the site — half-life and Tmax in the dosing visualizer, stability windows in the stability reference — carry a tier label that flags how strong the underlying source is. The convention is shared so a reader who learns it once can read it everywhere.
- T1 — peer-reviewed primary literature.
- T2 — regulator-reviewed labels (FDA approved drug labels and equivalents).
- T3 — manufacturer Certificate of Analysis or lab data published by the producer.
- T4 — applied USP/ICH or class-level standards used as a fallback when no compound-specific source exists.
Each charted or tabulated value displays its citation and tier next to the number, so the strength of the underlying source is visible alongside the value itself.
Half-life visualizer
The half-life visualizer plots a single dose of a peptide as a serum-level curve over time. The curve is normalized to its own peak (0–100%), not to an absolute concentration — bioavailability and volume of distribution are rarely published for peptides, and we do not guess them.
The math is a one-compartment Bateman model with first-order absorption: given the elimination half-life T½ and the time to peak Tmax from the cited source, we derive the elimination rate constant kel = ln(2) / T½ and solve for the absorption rate constant ka from the standard Bateman Tmax relationship Tmax = ln(ka/kel) / (ka − kel). The curve is C(t) ∝ e^(−kel·t) − e^(−ka·t), shown from t = 0 to ~5 × T½, with the x-axis unit (min / hours / days) chosen to fit that span.
Each charted peptide displays its citation and source tier (T1–T4) under the chart. See Source tiers for the convention.
Single-compartment is a simplification: many peptides exhibit multi-compartment distribution or non-linear elimination, and route of administration affects the absorption constant. This tool visualizes the shape of one published dataset under one model. It does not recommend a dose, a schedule, or a protocol.
COA reader
The COA reader accepts a public URL pointing to a Certificate of Analysis (PDF, PNG, or JPEG) and reports six standard fields from the document — lot number, issuing lab, purity (verbatim), identity test result, report date, and analyst or signer. (The extraction step also captures the tested compound name, but it is not surfaced in this release.) The tool is observational. It does not assess whether values are good, the document is authentic, or the product matches its label.
The pipeline has two stages. The LLM stage (Claude Sonnet 4.6) extracts verbatim strings or returns null when a field is not located — it is instructed to never infer, normalize, or score. The checklist stage is a deterministic function: a field is “present” if and only if its extracted value is a non-empty trimmed string. The LLM never decides presence; we do.
What we do not do: match the tested compound against catalog aliases, grade purity against thresholds, verify document authenticity, track expiry, or reconcile across multiple documents. Each row in the readout shows the verbatim extracted string with a per-row caveat to verify against the source.
Public access is rate-limited to 5 reads per IP per UTC day. Client IPs are hashed with HMAC-SHA-256 before storage; raw IPs are never persisted.
Limitations
- We do not provide dosage, protocol, or treatment recommendations.
- We do not test products ourselves; third-party verification reflects records we have aggregated, not in-house lab work.
- We do not verify the identity of the buyer, the legality in your jurisdiction, or your specific use case.
- We do not endorse any vendor. A vendor appearing on this site is not a recommendation to purchase.
- We do not capture every vendor. Tracked vendors are listed on the scraper policy page.
Methodology changelog
Material changes to the methodology are recorded here in reverse-chronological order.
Editorial scope FAQ published at /faq. States the catalog inclusion criteria, the four-vendor selection rules, the T1–T4 source-tier framework, and the suggest-a-peptide / data-correction flow in plain English.
Trust pages restructured. Methodology, About, Affiliate, Scraper Policy, and Contact moved to a shared layout with per-vendor scrape-cadence table sourced from one module.
Data-freshness language reconciled across the site so methodology, scraper policy, and listing pages describe vendor refresh cadence in the same words.
Transparency scoring entered beta. Vendor pages surface a Complete / Partial / Pending / Insufficient data status instead of a single composite score.
Methodology v1 published. Default sort by price-per-mg, in-stock-first, with affiliate status decoupled from rank.