Vercel AI SDK (Anthropic provider)

@ai-sdk/anthropic with the ai package, on plugboard's Messages API.

Connect

The provider's base URL ends in /v1; it appends /messages:

const baseURL = process.env.PLUGBOARD_BASE_URL! // ends in /v1
const apiKey = process.env.PLUGBOARD_KEY!

export const anthropic = createAnthropic({ baseURL, apiKey })

Chat

const { text, finishReason, usage } = await generateText({ model: make(model), prompt: "Reply with exactly: pong", maxOutputTokens: 1024 })

Tool calls

The SDK runs execute and sends the follow-up request. Plugboard only returns the call.

const get_weather = tool({
  description: "Current weather for a city",
  inputSchema: z.object({ city: z.string().describe("City name") }),
  execute: async ({ city }) => {
    cities.push(city)
    return weatherResult
  },
})
const result = await generateText({ model: make(model), tools: { get_weather }, stopWhen: isStepCount(4), prompt: PROMPT, maxOutputTokens: 1024 })

Structured output

Output.object works with the default structuredOutputMode (which sends output_config.format for Claude models) and with "jsonTool" (a forced tool named json). The example uses "jsonTool":

const { output } = await generateText({
  model: make(model),
  output: Output.object({ schema: Meal }),
  prompt: PROMPT,
  maxOutputTokens: 1024,
  providerOptions: { anthropic: { structuredOutputMode: "jsonTool" } },
})

Limits

• The SDK has no model-list call. GET /v1/models accepts the same key as x-api-key or Authorization: Bearer.
• Errors carry error.type and no error.code.
• The schema in structured output is not enforced by the provider. Validate the answer.