What is Inference Gateway?

Inference Gateway is Canvas's endpoint for Claude models. Your developers call it with a Canvas key, which your organization creates and manages in Canvas Platform.

Do we need our own Anthropic account?

No. Your developers use a Canvas key in place of an Anthropic key.

Does Inference Gateway cost extra?

Yes. Canvas bills Inference Gateway per token as an additional charge on your monthly Canvas invoice. Billing admins can see the per-model token prices on the Billing settings page in Canvas Platform.

Which models can we use?

The gateway serves Claude models through the Anthropic Messages API. Canvas can limit an organization to specific models.

Which Anthropic APIs does the gateway support?

The gateway serves message requests and token counting. Other Anthropic endpoints, such as Message Batches, Files, Skills uploads and Models, return an error. The priority service tier isn't available, so leave service_tier unset or use the standard tier.

Can we use Anthropic's server-side tools, such as web search?

Yes. The gateway passes tool definitions and tool calls inside a message request through unchanged. Canvas bills web search per search on top of tokens, and those charges appear in the same usage view.

Can we cap what we spend?

Yes. Your organization sets a monthly budget, and each key can have its own monthly limit. By default, reaching the budget only sends an alert. A Billing admin can turn on Stop requests instead.

Stop requests halts every workflow that uses your organization's keys, including patient-facing workflows, so check what depends on inference before you turn it on.

Does Studio use Inference Gateway?

Yes. Canvas Studio, the AI plugin builder for Canvas, makes its model calls through Inference Gateway on your organization's account. Token fees apply to Studio usage under your AI Addendum.

Can we send PHI through Inference Gateway?

Yes, once your organization signs the AI Addendum and your inference account is active. Requests go only to model providers that Canvas's BAA and zero data retention policy cover. Your team does not need its own BAA with the model provider for these calls.

Does Canvas store our prompts or the model's responses?

No. The gateway passes each request and response through unchanged and reads only the token counts it needs for billing. The gateway does not log or save prompts and completions. Usage records hold the key, the model and the token counts, which is what you see on the Usage & Billing page.

How is this different from using our own Anthropic or other model provider account?

The model is the same Claude you would reach directly, and the API is the same Anthropic Messages API. The differences are in everything around the model call.

With your own provider account, your organization has to put its own BAA and zero data retention terms in place with that provider before you send any PHI. Through the gateway, Canvas's BAA with Anthropic (and your BAA with Canvas) and your AI Addendum already cover those terms.

Gateway usage appears on your monthly Canvas invoice next to the rest of your Canvas usage. Every key can carry its own optional monthly limit on top of an organization budget, with email alerts and an optional hard stop, and you can see spend per key.

Your Organization admins create and revoke keys in Canvas Platform, so they can revoke a leaked key in one step. Studio already runs on your organization's gateway account, so your Studio usage and your own builds share a single budget and one usage view.

Do Claude Code skills, plugins and MCP servers work with a Canvas key?

Yes. They run in Claude Code on your machine or in your app and reach the model through ordinary message requests, so they work the same way through the gateway as with a direct Anthropic key. This includes the Canvas Plugin Assistant, a Claude Code plugin Canvas publishes for building Canvas plugins, with the Canvas SDK reference, plugin patterns, security review, test authoring and deployment workflows built in. Install it in Claude Code with /plugin marketplace add canvas-medical/coding-agents and then /plugin install cpa@canvas-medical.

The BAA covers the model providers behind the gateway. An MCP server or other tool that Claude Code calls is a separate service, so make sure your organization has its own agreement with any tool that receives PHI.

How do we get started?

Contact your Canvas team to sign the AI Addendum and turn on your inference account. The setup guide in the Canvas Help Center covers creating a key and connecting your code.