Audit log

The list is the product; this is the receipt of how it changes.

2,376 total entries id: privatemode × · JSON Feed · All entries (JSON)

• 18d ago policy_change provider privatemode
  details

  {
    "field": "fees_page",
    "summary": "Gemma 4 31B model (€0.77 input/€1.27 output/€0.08 cached per 1M tokens) removed from pricing table and model comparison list. Supported model lineup changed, affecting available services and pricing.",
    "url_source": "discovered_conventional"
  }

• 60d ago policy_change provider privatemode
  details

  {
    "field": "fees_page",
    "summary": "Two AI models removed from pricing table: Gemma 3 27B and Qwen3-Coder 30B-A3B no longer listed. Compare-plans also drops both models and bumps Kimi-K2.5→K2.6. Affects available inference options for p",
    "url_source": "discovered_conventional"
  }

• 77d ago update provider privatemode
  details

  {"provider":{"id":"privatemode","category":"ai","subcategory":"Confidential Compute","name":"Privatemode","tagline":"Confidential-compute LLM proxy — encrypted inference on Nvidia confidential GPUs.","review":"Runs the inference inside an attested TEE so even the operator can't see prompts. Open-source clients; pay-per-token, crypto accepted. Closest the LLM space gets to \"actually private\" right now.","kyc_tags":["anonymous_signup"],"feature_tags":["open_source","api_available"],"rating":{"stars":null,"count":0,"grade":"A"},"fees":{"summary":"Per-token · BTC / fiat · confidential-compute backend"},"links":{"web":"https://www.privatemode.ai","outbound_slug":"privatemode"},"highlight_tags":["CONFIDENTIAL","TEE"],"last_verified":"2026-05-12","_seed_i18n":{"zh":{"tagline":"在 Confidential Computing 環境中跑的 AI — 連 hosting 商也看不到內容。","review":"Privatemode 把 AI 推論放進 AMD SEV / Intel TDX 的可信執行環境，宣稱連雲端商自己都解密不了 prompt。透過硬體 attestation 證明執行緒在加密 enclave 內。對於必須使用雲端 AI 但又不能讓服務商讀到內容的合規敏感場景（法律、醫療、企業）有實際