Vercel AI Gateway for AI agents

Vercel AI Gateway routes requests to different AI models through one endpoint, used by developers who want model flexibility and usage tracking without juggling separate API keys. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.

AI and modelsAPI keyMCP + RESTvercel.com
AUDIT LOG · VERCEL AI GATEWAYPOLICY: acme-support
09:41:07 · claude · u_8f2read
vercel_ai_gateway.list_model_catalog
List Model Catalog✓ allowed · 212ms
09:41:08 · claude · u_8f2read
vercel_ai_gateway.get_model
Get Model✓ allowed · 164ms
09:41:09 · claude · u_8f2write
vercel_ai_gateway.create_embeddings
Create Embeddings✓ approved · approved by user
EVERY VERCEL AI GATEWAY CALL, ON THE RECORD
01 · USE CASES

What agents do in Vercel AI Gateway.

01

Create a chat completion

Send a prompt through the gateway to a chosen model and get back a response.

02

Check the spend report

Read usage and spend across models as a read-only report.

03

List the model catalog

Pull the available models before deciding which one to route a request to.

02 · ACTIONS

22 Vercel AI Gateway actions, graded by risk.

Every Vercel AI Gateway action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.

read

9

Look things up. Allowed by default.

  • vercel_ai_gateway.get_model
    Get Model
  • vercel_ai_gateway.search_models
    Search Models
  • vercel_ai_gateway.get_generation
    Get Generation
  • vercel_ai_gateway.get_spend_report
    Get Spend Report
  • vercel_ai_gateway.get_credit_balance
    Get Credit Balance
  • vercel_ai_gateway.list_model_catalog
    List Model Catalog
  • vercel_ai_gateway.get_model_endpoints
    Get Model Endpoints
  • vercel_ai_gateway.list_typesafe_models
    List TypeSafe Models
  • vercel_ai_gateway.count_message_tokens
    Count Message Tokens

write

13

Create and change things. Allow, or ask the user first.

  • vercel_ai_gateway.create_embeddings
    Create Embeddings
  • vercel_ai_gateway.create_chat_completion
    Create Chat Completion
  • vercel_ai_gateway.create_anthropic_message
    Create Anthropic Message
  • vercel_ai_gateway.edit_images
    Edit Images
  • vercel_ai_gateway.evaluate_state
    Evaluate State
  • vercel_ai_gateway.generate_images
    Generate Images
  • vercel_ai_gateway.generate_speech
    Generate Speech
  • vercel_ai_gateway.compact_response
    Compact Response
  • vercel_ai_gateway.transcribe_audio
    Transcribe Audio
  • vercel_ai_gateway.generate_response
    Generate Response
  • vercel_ai_gateway.typesafe_evaluate
    Evaluate with TypeSafe API
  • vercel_ai_gateway.rerank_text_documents
    Rerank Text Documents
  • vercel_ai_gateway.rerank_object_documents
    Rerank Object Documents

destructive

0

Delete, cancel or archive. Ask first, or deny outright.

  • No destructive actions.
03 · HOW IT WORKS

Vercel AI Gateway in three steps.

  1. 01Your users connect Vercel AI GatewayThey add their Vercel AI Gateway api key on Arc0 Connect, under your brand. It goes straight into the vault.
  2. 02You set the rulesReads run, and writes like “create Embeddings” can wait for the user to approve.
  3. 03Any agent can actYour agent calls Vercel AI Gateway through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
POLICY.TS
await arc0.policies.set('vercel_ai_gateway', {
  read: 'allow',
  write: 'ask',        // create_embeddings
  destructive: 'deny',  
})

# Claude Code: the same connection, one URL
$ claude mcp add --transport http arc0 \
    https://mcp.arc0.ai/u/u_8f2
04 · AUTH AND DATA

How Vercel AI Gateway connects.

Users add their Vercel AI Gateway api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

The same Vercel AI Gateway connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →

AUTH
API key
CREDENTIALS
Per-tenant encrypted vault
MODEL SEES
Results only, never credentials
AUDIT LOG
Every call, on every plan
05 · WORKS WITH

Use Vercel AI Gateway from any agent.

Claude
ChatGPT
Cursor
Codex
VS Code
OpenAI Agents SDK
Claude Agent SDK
Vercel AI SDK
Mastra
LangGraph
07 · FAQ

Vercel AI Gateway and Arc0, answered.

Q01

Can I use Vercel AI Gateway with Claude, ChatGPT or Cursor?

Yes. Connect Vercel AI Gateway to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Vercel AI Gateway actions you allow.

Q02

How do users connect Vercel AI Gateway?

Users add their Vercel AI Gateway api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

Q03

Which Vercel AI Gateway actions can my agent take?

22 in total: 9 read, 13 write and 0 destructive, such as “create Embeddings”. Your policies decide which of them each agent may call.

Q04

Can I make my agent read-only in Vercel AI Gateway?

Yes. Allow read actions and deny writes in the Vercel AI Gateway policy. Your agent can still look things up, and any write it attempts is blocked and logged.

Q05

Can my own backend call Vercel AI Gateway too?

Yes. The same Vercel AI Gateway connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.

Get started

Plug Vercel AI Gateway into your agent.

Your users connect Vercel AI Gateway once, under your brand. Your agent gets 22 actions behind your policies, with every call on the record.

Free to build · MCP + REST · Audit log on every plan