Observed arrival · 2026-09-13
Veridrop's AI API Middleman Truth榜
Credibility concern recorded. The source reference remains available for verification and correction.
A Chinese-language service that ranks AI API intermediaries by claimed model authenticity, uptime, latency, and protocol evidence.
Field notes
The service separates protocol-specific checks: Claude is evaluated through server-side thinking_signature validation, while OpenAI and Gemini entries are assessed through fields, usage data, compatibility, and behavioral patterns. The homepage reports 23 current Astra checks and 11 Fable 5.1 checks, alongside intermediary-level uptime, latency, and detection counts. Operators authorize the API keys used for testing, while the underlying keys and raw evidence remain private.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue