LIVE MONITORING
AI API Relay Checker
Catch diluted quotas, model downgrades and rebranded wrappers before they cost you. Enter your endpoint for a free, instant availability and latency check.
🧬 How do you tell if a relay is downgrading or rebranding models?
A seconds-long probe of whether the relay works at all: one minimal request, then connectivity, latency and the raw reply — over OpenAI, Claude or Gemini protocols.
What is AI API relay dilution?
AI API relay dilution is when a relay provider silently swaps the model you asked for with a cheaper one (rebranding), cuts the context window or reasoning budget (downgrading), or fabricates responses outright. CCDetect (AI API relay fidelity checker) sends a set of forensic probes to the endpoint to catch this: it compares tokenizer fingerprint deltas, checks whether thinking signatures are well-formed, and verifies that the model's self-report matches its identity echo. A deep check runs the full probe suite in about five minutes; a connectivity check returns availability and latency in seconds. OpenAI-compatible, native Claude (/v1/messages) and native Gemini protocols are all supported, unregistered users get 10 free checks per hour, and your API key is used for that request only — never stored, never logged.
⭐ Which AI API relays are worth using?
See all listed relays →
绝不掺水的纯粹算力源
Featured slots include paid placements; payment affects ordering only, never credit scores. Credit grades still reflect the public record, and a listing is not an endorsement — do your own due diligence.