Guides
Guides
The documentation describes what each part of Salyro does. A guide works through one whole task instead — from the first step to a result you can check.
- Debug a failed requestAn ordered path from "something failed" to "this is why" — the status, the error body, the request id, and the gateway's log.
- Migrate an existing OpenAI integrationMoving production traffic onto Salyro is a configuration change — a base URL, a key and a model id — not a rewrite.
- Send images with visionThe second modality Salyro supports — how an image goes into a request, which providers read it, and what it does to token counts.
- Stream responsesShow text as it is generated instead of waiting for the whole answer — enabling streaming, consuming it server-side, and handling a client that leaves.
- Get structured JSON outputAsk for output that conforms to a schema instead of extracting JSON from prose — and find out immediately when a model cannot do it.
- Switch providers without changing your codeWith one interface in front of four providers, moving a request from OpenAI to Anthropic is a change of model id — and nothing else.
- Attribute cost per feature or customerSplitting gateways along the boundary you want to measure, so “what did this feature cost us” has an answer.
- Respond to a leaked API keyWhat to do in the first few minutes after a Salyro key reaches somewhere it should not have — and what you can establish afterwards.
- Rotate a provider credential without downtimeThe order of operations that replaces a provider key on a gateway without a window where requests fail.
- Separate your environments with gatewaysHow to give development, staging and production a gateway each, and what that separation actually buys you.
