Editorial Disclosure
Our position made public. Relationships with USDT card issuers, ad partners, and other stakeholders — so readers can evaluate our objectivity for themselves.
usdtcard.net is an independently operated USDT virtual card review and aggregation platform. This page discloses every relationship we have with card issuers, advertisers, and other stakeholders.
What we don't do
- Do not issue cards or act as issuer agent
- Hold no equity in any card issuer
- Do not accept issuer payment to raise scores, rewrite reviews, or pre-publish removal
- Do not preview scores or rankings to issuers before publication
- Do not accept paid promotional content ("soft articles") from crypto projects
Revenue sources
The site currently has no commercial revenue. Future revenue plans (with policy updates on this page when activated):
- Affiliate links — Some issuers may offer referral links. If activated, each card review will explicitly mark the link with
★ Affiliateand we guarantee: affiliate links do not influence the card's score or ranking. - Display advertising — No plans. If introduced, will be restricted to non-crypto brands and clearly marked
Sponsored. - Premium subscription — No plans. If launched, paid features will be limited to "advanced filters / API access / data export" type tooling, never gating editorial content.
Issuer relationships
We have no commercial relationship with any reviewed issuer at this time. All fee data is sourced from issuer public web pages, linked at the bottom of each card review.
Editorial process
- Scoring formula is public — see methodology
- All raw data sourced from issuer official pages, linked at the bottom of each card review
- Score changes exceeding ±0.5 require human review before publication
- Card SKU status changes (live ↔ unavailable) require human confirmation
AI assistance boundary
We use Anthropic Claude Opus model to assist with some content. Readers have the right to know the real source of each piece, so we explicitly disclose:
- AI-assisted generation: Daily industry news (
/news/*) is drafted by Claude Opus from a topic queue, then passed through three guardrails (link health check + Claude self-rating + content blocklist) before publishing. These pieces are markedauthor.name: "usdtcard editorial"in frontmatter and footer-tagged "AI-assisted". - Human-written: All card reviews (
/cards/*), best listicles (/best/*), beginner guides (/guides/*), scenario walkthroughs (/scenarios/*), and compliance analysis (/compliance/*) are written by human editors. AI does not participate in scoring or judgment for these. - Data collection: Hourly automated scrape of issuer official fee pages (no AI involved). Structured data goes directly into the database.
- Guardrail mechanism: AI-generated content must pass before publication: (1) all external links return HEAD 200; (2) Claude self-rates 4 dimensions (factual accuracy / readability / originality / compliance risk) ≥ 8/10 total; (3) content blocklist (no "guaranteed return", "100% safe", etc.); (4) daily publish cap = 5. Any failure → draft PR queued for human review.
Editorial governance
- Repository code and content live in private Git; all publishing flows through GitHub Actions → Cloudflare Pages
- Currently operated by 1 editor-in-chief; commit and push privileges are not delegated
- Autopilot pipeline runs under a separate account (
[email protected]) with content-file commit privileges only — no rating modification or review authoring - Every published version is traceable to a specific commit; corrections are logged on this page
Audit and review cadence
- Every Monday: top picks recompute based on the week's fee + status data
- Tuesdays and Fridays: scenario walkthrough refresh
- 1st of each month: full compliance page review (by region)
- Quarterly: methodology scoring formula audit; adjustments documented if any
- Annual: transparency report (total revenue / total corrections / total complaints)
How to report issues
If you find any content factually incorrect, outdated, or influenced by a third party, email [email protected]. We commit to: 48-hour response, 72-hour correction once verified, and public disclosure of the correction in the log below.
Correction log
None to date.