Articles
Guides to routing, reviewing and paying for AI inference, from the team behind Inference Bureau.
Gateway
Putting a model in a Shopify app without a second API bill
An app's actions call models through the ai client: ordinary gateway requests under your Directives and budgets, with their cost in the app's Logs.
GatewayWhat an AI code review costs: Auditor's model for each role, priced
Auditor bills each review as ordinary model requests. The default model for each role, its price, the per-review math and what our eval cases cost.
GatewayImage generation prices compared: Recraft, Seedream and Grok Imagine
Per-image prices for every Recraft, Seedream and Grok Imagine model, from $0.007 to $0.30 an image, and how to call them all with one OpenAI SDK.
GatewayWhy tool calls fail on some providers of the same open model
The same open model can return clean tool calls on one provider and broken JSON on another. What probing 280 routes showed, and how routing avoids the bad ones.
Host apps and functions next to your model traffic. Routing picks each turn's model by your quality floor, and Auditor reviews every pull request against your own rules.