公开 API
只读 JSON 接口,数据与站点展示的结算记录同源。无需密钥、无需注册,合理使用不限速。发布数据时需署名并回链。
| Endpoint | 这个端点回答什么 | 参数 |
|---|---|---|
GET/api/v1/coins
{"as_of":"2026-07-24","coins":[{"symbol":"BTC","stance":"neutral","grades":{...}}]} |
Every coin on the docket with its current verdict and grades — the whole board in one call. | — |
GET/api/v1/coin/{symbol}
{"symbol":"BTC","verdict":{"stance":"neutral","headline":"..."},"dimensions":{...}} |
Today's full research verdict for one coin, including every graded dimension. | symbol — ticker, e.g. BTC |
GET/api/v1/coin/{symbol}/{vdate}
{"symbol":"BTC","as_of":"2026-07-17","verdict":{...}} |
The archived verdict for one coin on one past day — what we said back then, unedited. | symbol — ticker; vdate — YYYY-MM-DD |
GET/api/v1/receipts
{"board":{"W":12,"L":9,"P":5,"decided":21,"winrate":"57%"},"total":635,"rows":[...]} |
The settled record: every call graded against the market, with both price endpoints. Page through the whole ledger with limit/offset — total tells you how many rows there are. | limit — rows per page, default 200 (≤1000); offset — rows to skip, default 0; response includes total |
GET/api/v1/kol
{"ranked":[{"author":"...","hit_rate_pct":62.5,"n_verified":48,"n_prospective":48,"hits_prospective":30,"prospective_rate_pct":62.5,"n_retrospective":22,"hits_retrospective":9,"retrospective_rate_pct":40.9,"n_total":70}],"building":[...]} |
KOL scoreboard — hit rate and sample size per account, from the settled call ledger. Prospective and backtested calls are scored on separate lines and never averaged together; ranking eligibility uses n_total. Legacy keys (n_verified, hits, hit_rate_pct) remain the prospective line. Behavior change (ruling E, 2026-07-30): ranked membership/order now follows n_total, so retro-only accounts may appear with n_verified=0 and null rates. Caliber change (2026-08-16): rates are graded under kol_directional_v1 — direction held over a 7-day window with a ±3% band, altcoins scored vs BTC and BTC vs USD, identical parameters to the settle_v1 rule used on CoinVerdict's own verdicts. The top-level `rule` key names the caliber the numbers were graded under; price_target_v1 there means the board fell back to price-target records. n_pt_prospective is the sample size of the OTHER yardstick (price_target_v1) and must never be added to n_total — two rules summed give a percentage nobody computed. | — |
GET/api/v1/commits
{"commits":[{"utc_date":"2026-07-24","chain_sha256":"...","n_reports":100}]} |
Daily cryptographic commitment chain — proof that published records were not edited later. | — |
GET/api/verdict/{symbol}
{"status":"done","report":{...}} |
Cached verdict for one coin, or the generation status if today's run is still pending. | symbol — ticker, e.g. BTC |
GET/api/price/{symbol}
{"symbol":"BTC","price_usd":64928.0,"sources":5} |
Median spot price across five exchanges — the same snapshot the verdicts are graded on. | symbol — ticker, e.g. BTC |
GET/api/healthz
{"ok":true} |
Liveness probe for uptime monitoring. | — |