Vercel AI Gateway

VercelAI & MLFree tier available

Managed gateway for calling models from many AI providers through one API (AI SDK, OpenAI- and Anthropic-compatible), with centralised credentials, BYOK, request logs, budgets, routing and provider/model failover, and no markup on provider token prices

FluffyStack tools
Add to Service BuilderAdd to CompareCompare with equivalentsExplore Vercel in TreemapExplore ai-ml in HoneycombSee Vercel regions on the World MapSee ai-ml as a networkScore jurisdiction exposure
DocumentationPricingVercel website

Jurisdictional exposure

Provider HQ
USSan Francisco, USA

Subject to CLOUD Act, FISA-702, DPF

Region locations
APACCNEUUKUSOther19 regions across 6 jurisdictions
Sovereign option
No sovereign-flagged regions in the catalogue for this service.
Full scorecard for this service →US lens detail →Sovereign cloud coverage map →

Sub-services (4)

Bring Your Own Key

Use existing provider API keys with no added fee

Provider and model fallbacks

Automatic retry on alternative providers or models when a request fails

Observability

Request logs, token usage and spend per model and project

OpenAI- and Anthropic-compatible APIs

Drop-in endpoints for existing SDKs alongside the AI SDK

Compliance & Certifications

This service is attested for the following frameworks. Always verify with the provider before relying on a specific compliance posture.

GDPRSOC 2ISO 27001HIPAA

Where this runs

19 regions
15 countries
Commercial regions (19)

Europe (5)

  • Paris, France
  • Frankfurt, Germany
  • Dublin, Ireland
  • Stockholm, Sweden
  • London, UK

North America (5)

  • Montréal, Canada
  • Cleveland, USA
  • Portland, USA
  • San Francisco, USA
  • Washington, D.C., USA

South America (1)

  • São Paulo, Brazil

Asia (6)

  • Hong Kong
  • Mumbai, India
  • Osaka, Japan
  • Tokyo, Japan
  • Singapore
  • Seoul, South Korea

Oceania (1)

  • Sydney, Australia

Africa (1)

  • Cape Town, South Africa

Tags

ai-gatewayllmmodel-routingbyokfailover

Equivalent services on other platforms

Cloudflare AI GatewayCloudflare

Unified proxy layer for LLM API traffic with caching of identical prompts, per-provider rate limits, retry and fallback policies, cost and latency analytics, and prompt logging for audit — works across OpenAI, Anthropic, Google, and Workers AI

Pricing

Pricing model:per-token-pass-through