name	vercel-ai-gateway
description	Vercel AI Gateway expert guidance. Use when configuring model routing, provider failover, cost tracking, or managing multiple AI providers through a unified API.
metadata	{"author":"Vercel","source":"https://github.com/vercel/vercel-plugin","upstreamPath":"skills/ai-gateway/SKILL.md","priority":7,"docs":["https://vercel.com/docs/ai-gateway","https://sdk.vercel.ai/docs/ai-sdk-core/settings"]}

Vercel AI Gateway

Before writing gateway code, fetch the current Vercel AI Gateway docs and provider/model list. Model slugs, routing rules, and authentication behavior change frequently.

Core usage

Use provider/model format, for example openai/gpt-5.4 or anthropic/claude-sonnet-4.6. Prefer discovering available model IDs before hardcoding them.

import { generateText } from 'ai'

const result = await generateText({
  model: 'openai/gpt-5.4',
  prompt: 'Hello!',
})

Use gateway() only when you need explicit gateway options such as provider order, failover, tags, or per-user routing.

import { generateText, gateway } from 'ai'

const result = await generateText({
  model: gateway('anthropic/claude-sonnet-4.6'),
  providerOptions: {
    gateway: {
      order: ['anthropic', 'bedrock'],
      models: ['openai/gpt-5.4'],
      user: 'user-123',
      tags: ['feature:chat', 'env:production'],
    },
  },
  prompt: 'Hello!',
})

Model slug rules

Use provider/model format.
Versioned model slugs use dots, not hyphens.
Do not default to outdated models without checking the docs.
Prefer AI Gateway for production apps, multi-provider failover, observability, cost attribution, and per-user rate limiting.

Authentication

For Vercel projects, prefer OIDC setup:

vercel link
vercel env pull .env.local

For CI or non-Vercel environments, use AI_GATEWAY_API_KEY when OIDC is not available.

Error handling

Handle budget and rate-limit errors explicitly:

HTTP 402: budget or hard spending limit exceeded.
HTTP 429: rate limit exceeded.
HTTP 503: provider unavailable.

Official documentation

Source

Condensed from vercel/vercel-plugin/skills/ai-gateway/SKILL.md. The upstream file is larger; use the upstream repo for the full executable Claude plugin behavior.

Vercel AI Gateway

Before writing gateway code, fetch the current Vercel AI Gateway docs and provider/model list. Model slugs, routing rules, and authentication behavior change frequently.

Core usage

Use provider/model format, for example openai/gpt-5.4 or anthropic/claude-sonnet-4.6. Prefer discovering available model IDs before hardcoding them.

import { generateText } from 'ai' const result = await generateText({ model: 'openai/gpt-5.4', prompt: 'Hello!', })

Use gateway() only when you need explicit gateway options such as provider order, failover, tags, or per-user routing.

import { generateText, gateway } from 'ai' const result = await generateText({ model: gateway('anthropic/claude-sonnet-4.6'), providerOptions: { gateway: { order: ['anthropic', 'bedrock'], models: ['openai/gpt-5.4'], user: 'user-123', tags: ['feature:chat', 'env:production'], }, }, prompt: 'Hello!', })

Model slug rules

Use provider/model format.

Versioned model slugs use dots, not hyphens.

Do not default to outdated models without checking the docs.

Prefer AI Gateway for production apps, multi-provider failover, observability, cost attribution, and per-user rate limiting.

Authentication

For Vercel projects, prefer OIDC setup:

vercel link vercel env pull .env.local

For CI or non-Vercel environments, use AI_GATEWAY_API_KEY when OIDC is not available.

Error handling

Handle budget and rate-limit errors explicitly:

HTTP 402: budget or hard spending limit exceeded.

HTTP 429: rate limit exceeded.

HTTP 503: provider unavailable.

Official documentation

Source

Condensed from vercel/vercel-plugin/skills/ai-gateway/SKILL.md. The upstream file is larger; use the upstream repo for the full executable Claude plugin behavior.

vercel-ai-gateway