Skip to main content
Your Cloud key is the model credential. Leave the models slot unset and every turn rides the console’s gateway.

​
How the gateway is wired

Managed inference is not a bespoke client. It is the stock @ai-sdk/anthropic provider, pointed at the console instead of at Anthropic.

One provider, one base URL swap. VENDO_API_KEY is the bearer token.

The console speaks the Anthropic Messages wire, which is the whole reason the Anthropic provider serves it. Point the same provider at https://console.vendo.run/api/v1 and it works.
Gateway traffic does not carry Vendo’s deployment-identity headers, so inference does not appear in the console’s deployment inventory. Usage still meters.

​
Which model you get

The gateway serves the vendo model family as literal model ids. The console maps each name to a concrete model server-side, so a Cloud-keyed app can be retuned without a client release. Names pass through verbatim. There is no client-side translation, so an unknown name surfaces the gateway’s own error rather than a Vendo one. Leave a seat unset and it takes its own family id here, so an empty models block still puts each job on the name tuned for it. The default seat is also published in three billed tiers, vendo-fast, vendo, and vendo-strong. Their rates are on the Vendo Cloud page.

​
Pinning a model

Two ways, and an explicit model object always wins over both.
app/api/vendo/[...vendo]/route.ts
.env.local
Precedence runs top to bottom: an explicit model object, then the environment pin, then the configured string, then that seat’s own family name.

​
When there is no key

Every seat resolves through vendoModel(), lazily, on the first call that needs a model. With no credential that call fails, and the message names VENDO_API_KEY and vendo login as the way out. Nothing else in the environment fills a seat. A provider key sitting in your shell is a credential, not a selection, and Vendo never picks a provider on your behalf. If your product already runs its own agent loop, Vendo’s internal turns still need models of their own and will not borrow yours. Leaving the seats unset is what hands them the gateway.

​
Verifying the model

vendo doctor names the credential the runtime will resolve. It reads your environment and makes no model call.
Terminal
A key that is set but malformed reports E-CLOUD-001.

​
Where to go next

Vendo Cloud

What the key covers, what it costs, and where the rates are published.vendo login

Edge runtimes

The gateway spelled out by hand, for runtimes without Node.createAnthropic({ baseURL })

Environment variables

Every VENDO_* variable the runtime reads.VENDO_MODEL · VENDO_CLOUD_URL