AI Gateway

Fresh

by Vercel

LLM request proxy with provider routing, fallback chains, usage/spend tracking, BYOK, and OpenAI/Anthropic-compatible REST endpointsDevelopers building AI features who need multi-provider redundancy, cost visibility across models, and a single API key instead of juggling separate provider credentials and billingFree tier

Unified API proxy routing requests to 100+ LLM providers through a single endpoint with automatic fallbacks, load balancing, spend monitoring, and no per-token markup.

Features

Automatic Git-based CI/CD deploymentsall plansPer-branch and per-PR preview deploymentsall plansGlobal Edge Network (125+ PoPs, 20 Vercel regions)all plansIncremental Static Regeneration (ISR)all plansFluid Compute (serverless with server-process semantics)all plansEdge Middlewareall plansFeature Flagsall plansSkew Protection (deployment version pinning)paid onlyRolling Releasespaid onlyVercel Workflows (event-triggered automation pipelines)all plansOpenTelemetry tracing via CLI (vercel curl --trace)all plansSpend management and project-level budget capspaid onlyCustom WAF rules (up to 40 on Pro, 1,000 on Enterprise)paid onlyManaged WAF rulesetsenterprise onlyIP/CIDR blocking (up to 100 Pro, 1,000 Enterprise)paid onlySAML SSOpaid onlySCIM directory synchronizationenterprise onlyAudit logsenterprise only99.99% uptime SLAenterprise onlyMulti-region compute with automatic failoverenterprise onlyAI Gateway — zero-markup LLM proxy with 100+ modelsall plansAI Gateway BYOK (Bring Your Own Key)all plansAI SDK v5/v6 native integrationall plansThird-party integrations marketplaceall plansAdvanced Deployment Protection (password/Vercel auth gating)paid only

More from Vercel