Moving from Vercel AI Gateway?Your own keys, no credits to buy.

2 lines of code + 1 workflow

Keep your OpenAI SDK. On the client line, change the base URL and key. On the request line, set model to the name of a workflow.

Step 0: create a workflow in the console with your primary model and backups. Its name is what you send as model.

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://ai-gateway.vercel.sh/v1",
    api_key=os.environ["AI_GATEWAY_API_KEY"],
)

resp = client.chat.completions.create(
    model="anthropic/claude-sonnet-5",
    extra_body={"providerOptions": {"gateway": {"models": ["openai/gpt-4o"]}}},
    messages=[{"role": "user", "content": "Summarize this ticket."}],
)
print(resp.choices[0].message.content)

import os
from openai import OpenAI

client = OpenAI(
    base_url="https://api.modelriver.com/v1",
    api_key=os.environ["MODELRIVER_API_KEY"],
)

resp = client.chat.completions.create(
    model="support_bot",  # workflow: claude-sonnet-5 + 2 backups
    messages=[{"role": "user", "content": "Summarize this ticket."}],
)
print(resp.choices[0].message.content)

import { generateText } from "ai";

// AI_GATEWAY_API_KEY is read from the environment
const { text } = await generateText({
  model: "anthropic/claude-sonnet-5",
  prompt: "Summarize this ticket.",
  providerOptions: {
    gateway: { models: ["openai/gpt-4o"] },
  },
});
console.log(text);

import { generateText } from "ai";
import { createOpenAI } from "@ai-sdk/openai";

const modelriver = createOpenAI({
  baseURL: "https://api.modelriver.com/v1",
  apiKey: process.env.MODELRIVER_API_KEY,
});

const { text } = await generateText({
  model: modelriver.chat("support_bot"), // .chat() = Chat Completions
  prompt: "Summarize this ticket.",
});
console.log(text);

curl https://ai-gateway.vercel.sh/v1/chat/completions \
  -H "Authorization: Bearer $AI_GATEWAY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-sonnet-5",
    "providerOptions": {"gateway": {"models": ["openai/gpt-4o"]}},
    "messages": [{"role": "user", "content": "Summarize this ticket."}]
  }'

curl https://api.modelriver.com/v1/chat/completions \
  -H "Authorization: Bearer $MODELRIVER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "support_bot",
    "messages": [{"role": "user", "content": "Summarize this ticket."}]
  }'

What maps to what

Vercel AI Gateway concepts and where they live in ModelRiver.

Map your config

Your fallbacks, output shape and budget move into one named workflow.

{
  "model": "anthropic/claude-sonnet-5",
  "messages": [{ "role": "user", "content": "Summarize this ticket." }],
  "providerOptions": {
    "gateway": {
      "order": ["anthropic", "bedrock"],
      "models": ["openai/gpt-4o", "google/gemini-3.5-flash"]
    }
  }
}

When to stay where you are

Vercel AI Gateway may suit you better if any of these matter most.

• You want the AI SDK's built-in gateway: plain provider/model strings with no provider setup.

• You want one credit balance that pays for every provider.

• You export OpenTelemetry traces with trace drains.

• You need zero data retention routing, prompt-training blocks or regional inference.

• You use image, video, speech or reranking models, the Responses API or the Anthropic Messages API.

• You route one model across several providers with order, only or sort.

What you gain, what you give up

Steps to move over

1. Step 1

   Add your provider keys

   Add a key for each provider your models use. ModelRiver has no credits, so every model in your workflows needs a key. Keys are encrypted at rest and masked once saved.

   Vercel AI Gateway docs (opens in a new tab)ModelRiver docs

2. Step 2

   Turn each model setup or virtual model into a workflow

   Primary: the model you call today, or the model behind your vmc/ slug. Backups: the first two entries of your model fallbacks. Name it with letters, digits and underscores, such as support_bot.

   Vercel AI Gateway docs (opens in a new tab)ModelRiver docs

3. Step 3

   Compare the Draft with your current calls

   Run a few prompts you send to AI Gateway today in the Playground against the Draft. To wire up your app without provider calls, switch the workflow to Testing mode; switch back to Production before you publish.

   Vercel AI Gateway docs (opens in a new tab)ModelRiver docs

4. Step 4

   Publish the workflow Live

   API requests run the Live version, so publish before you switch code. The workflow name then plays the part of your vmc/ slug.

   Vercel AI Gateway docs (opens in a new tab)ModelRiver docs

5. Step 5

   Swap the provider, key and model

   In the AI SDK, create an OpenAI provider with baseURL: "https://api.modelriver.com/v1" and call .chat("support_bot"). With the OpenAI SDK, change the base URL, key and model, and remove providerOptions.

   const modelriver = createOpenAI({
     baseURL: "https://api.modelriver.com/v1",
     apiKey: process.env.MODELRIVER_API_KEY,
   });
   model: modelriver.chat("support_bot")

   Vercel AI Gateway docs (opens in a new tab)ModelRiver docs

6. Step 6

   Keep AI Gateway as a fallback for a while

   Keep your current gateway configured in your code as a fallback until you're confident, then remove it. With the AI SDK, keep the old provider/model string and use it if the ModelRiver call throws.

   Vercel AI Gateway docs (opens in a new tab)ModelRiver docs

7. Step 7

   Edit the workflow, not the virtual model

   Where you used to edit a virtual model, edit the workflow Draft, test it, publish it, and roll back to an earlier revision if the change misbehaves.

   Vercel AI Gateway docs (opens in a new tab)ModelRiver docs

Related

• Moving from LiteLLMMigrate from a self-hosted LiteLLM proxy to ModelRiver: hosted, OpenAI-compatible, your own provider keys. Map config.yaml fallbacks to one
• Moving from OpenRouterMigrate from OpenRouter to ModelRiver: OpenAI-compatible, your own provider keys, no token markup. Turn your models[] fallbacks into a versi
• Moving from PortkeyMigrate from Portkey to ModelRiver: move the fallbacks in your x-portkey-config header into one named workflow. Your own keys, flat per-requ
• All migration guidesWhat moves and what doesn’t, for each gateway.
• Gateway migration docsConcept mapping, snippets and first errors.