Can I override the default model for specific tasks?

Yes, the exported getModel() function accepts an optional modelId parameter, allowing you to use specialized reasoning models for complex tasks while keeping a lightweight default.

Do I need a Vercel account to use this skill?

While it is optimized for Vercel AI Gateway, it also supports local development with Ollama, which requires no external account or API keys for local inference.

What is the primary benefit of the AI Core skill?

It provides a standardized foundation for all AI features in your app, allowing you to manage model providers, environment variables, and local vs. cloud execution in one central place.

Which AI models are supported by this skill?

It supports over 200 models via Vercel AI Gateway (including Anthropic Claude, OpenAI GPT, and Google Gemini) and local models like Qwen via Ollama.

How does it handle environment variables?

The skill automatically updates your .env.local and src/env.ts files with the necessary keys and schemas, and can optionally pull secrets from 1Password.

AI Core Foundation

Name: AI Core Foundation
Author: mattwoodco

bymattwoodco

•

Data Science & ML

Configures a unified AI Gateway layer to manage and switch between 200+ large language models through a single interface.

AI Core serves as the foundational integration layer for Next.js applications using the Vercel AI SDK. It streamlines AI development by centralizing model configuration through Vercel AI Gateway, allowing developers to switch between providers like Anthropic, Google, and OpenAI without refactoring downstream code. The skill automates environment variable setup, integrates with Zod for schema validation, and provides a flexible utility that supports both cloud-based providers and local inference via Ollama for a seamless development-to-production workflow.

Key Features

01Secure API key retrieval via 1Password integration

02Automated environment variable configuration with Zod schema validation

03Seamless switching between cloud providers and local Ollama instances

043 GitHub stars

05Unified access to 200+ models via Vercel AI Gateway

06Centralized getModel() and getEmbeddingModel() utility functions

Use Cases

01Enabling local-first AI development using Ollama for privacy, speed, or offline testing

02Implementing cost-effective development workflows by defaulting to light models and swapping to reasoning models for production

03Standardizing AI model access across multiple downstream features or application skills

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mattwoodco/skills ai-core

For use in Claude.ai and ChatGPT

Download Skill