Guides & pitfalls
Practical material on AI API relay services: how rates are calculated, how to verify a model has not been downgraded, what enterprises should check before buying, and the pitfalls we ran into while measuring these services ourselves.
Browse by topic
All articles
Tool use compatibility across relay sites, and how to self-test it
Tool calling is the capability most likely to pass a test and fail in production. This article covers the structural dif…
Six kinds of streaming (SSE) failure on relay sites, and how to pin down each one
Streaming is where relay site quality shows through most clearly. This article lists six observable failure patterns, ea…
The hidden cost of self-hosting a relay: what it costs per month in money and time
Open source and free describes the price of the software, not of the server, the domain, or the most expensive part, mai…
How Sub2API works and where its compliance line sits
Projects like Sub2API turn a web subscription account into an API and reach a tenth of the official price. This article…
New API's protocol conversion: why it can be reached with Anthropic's official SDK
The dialect differences between vendors are the biggest technical pain point for relay services. This article explains w…
Choosing between One API and New API: an operations view, not a feature list
Feature comparison tables are everywhere, but what really decides which one suits you is who maintains it and whether yo…
Official resale, mixed routing, reverse-engineered: a three-way split of upstream sources and how to identify them
Two sites are both relay services, but different upstream sources give them entirely different stability floors and pric…
45.83% of endpoints failed identity verification: how to read that industry figure
The figure that 45.83% of endpoints failed identity verification is often used as proof that nearly half of relay sites…
Comparing review platforms for relay services: what apiranking, Ztest and GatewayBench each measure
There are more review platforms every year, but they are not measuring the same thing. This article breaks down the crit…
The 2026 relay site ecosystem in mainland China: three kinds of operator behind 2,000 sites
The number of relay sites in mainland China has passed 2,000, but by upstream source and operating model only three kind…
Where the compliance line sits when enterprise data goes through a third-party API
Where the real risk points are when an enterprise uses a third-party API, which practices reduce that risk, and when to…
Common error codes explained: what 401, 403, 404, 429 and 500 each mean
A lookup table you can consult directly: the common causes of each error code, who to contact, and what to check first.…
Supporting data & tools
The articles explain the method; the pages below give you measured data and prices to check it against. Reading them together is easier — once you know how to tell whether a rate is real, you can look the official price up in the price list.