所有目錄變更的公開分類帳——等級調整、KYC 重測、贊助啟用、事件更新。我們把它記下來,是為了讓訪客能驗證我們所說的「編輯防火牆」確實成立。
清單本身就是產品;這頁是它變動的收據。
{
"field": "tos",
"actor": "admin"
}{
"field": "privacy",
"actor": "admin"
}{
"field": "review",
"added": "policy update paragraph (policy-monitor 2026-11-08)"
}{
"field": "tos",
"summary": "Terms now state conversations are held in memory up to 1 hour before deletion, replacing the prior \"we do not store conversation content\" statement.",
"url_source": "discovered_footer"
}{
"field": "privacy",
"summary": "Prompt handling now says conversations are held in memory for up to 1 hour before deletion; the prior text said no local prompt cache.",
"url_source": "discovered_footer"
}{"provider":{"id":"obscurify","category":"ai","name":"Obscurify.ai","subcategory":"API · Multi-Provider","tagline":"Anonymous AI chat and an OpenAI-compatible API over 200+ models, with a random-number account and Monero vouchers.","review":"Obscurify gives you a chat interface and an OpenAI-compatible API in front of 200+ models. Guests can chat for free with a free model (Llama 3.3 70B when we looked); an account is just a random number, with no email. You pay with prepaid vouchers bought through their own BTCPay server, in Monero, bitcoin, Lightning or card, and the voucher isn't tied to your account.\n\nThere's an onion mirror, a warrant canary, and an option to run some models locally in your browser. The web client's source is on GitHub under the GPL, though the repository only went public in October 2026.\n\nThe limit is the usual one for proxies: your prompts go to the upstream model providers under Obscurify's credentials, so the provider doesn't see you, but Obscurify sees yo{
"name": "Obscurify.ai",
"category": "ai"
}