CtrlK
BlogDocsLog inGet started
Tessl Logo

ai-gateway

Vercel AI Gateway expert guidance. Use when configuring model routing, provider failover, cost tracking, or managing multiple AI providers through a unified API.

75

1.58x
Quality

66%

Does it follow best practices?

Impact

100%

1.58x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./skills/ai-gateway/SKILL.md
SKILL.md
Quality
Evals
Security

Evaluation results

100%

23%

Production Chat API with Provider Resilience

Multi-provider failover with error handling

Criteria
Without context
With context

provider/model format

100%

100%

Dots not hyphens in versions

0%

100%

gateway() wrapper with providerOptions

75%

100%

Anthropic primary provider

100%

100%

Azure failover in order

0%

100%

Fallback model via models array

100%

100%

Per-user tracking

100%

100%

Tags for cost attribution

100%

100%

No outdated default models

50%

100%

429 rate limit handling

80%

100%

402 budget exceeded handling

100%

100%

503 service unavailable handling

66%

100%

Correct package in package.json

100%

100%

100%

52%

Migrating from Direct Provider SDK to AI Gateway

OIDC auth setup and model routing

Criteria
Without context
With context

OIDC auth documented

0%

100%

No provider API keys

100%

100%

Removes @ai-sdk/anthropic import

100%

100%

Correct ai package version

0%

100%

provider/model format used

100%

100%

Dots not hyphens in versions

100%

100%

Up-to-date default model

100%

100%

No gateway() without providerOptions

0%

100%

VERCEL_OIDC_TOKEN mentioned

0%

100%

Token expiry addressed

0%

100%

100%

36%

AI Feature Cost Management for a Multi-Tenant SaaS

Cost tracking caching and budget controls

Criteria
Without context
With context

Cache for clause extraction

30%

100%

No cache for chat assistant

100%

100%

Per-customer user tracking

50%

100%

Feature tags on requests

37%

100%

Pre-flight token estimation

100%

100%

402 budget error handled

70%

100%

streamText for summarization

100%

100%

Dots in model version slugs

0%

100%

No outdated gpt-4o model

0%

100%

usage field logged

100%

100%

Repository
vercel/vercel-plugin
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.