Add Provider Package

When you need to integrate a new AI provider into the Vercel AI SDK as an official @ai-sdk package, this walks you through the entire setup. It covers the layered architecture (specifications, utilities, providers, core), shows you how to structure your package directory, and explains the adapter pattern for implementing LanguageModelV4, EmbeddingModelV4, and other model types. The guide is thorough on the details that matter: using workspace dependencies, configuring both Node and Edge tests, handling API keys with loadApiKey, and the critical distinction between user-facing schemas (use .optional) and API response schemas (use .nullish). It also covers the less obvious stuff like running pnpm update-references after setup and why you shouldn't create a CHANGELOG.md file yourself.

skills.sh installs: 1.4K
GitHub stars: 27.8K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add vercel/ai --skill add-provider-package --agent claude-code

Installs into .claude/skills of the current project.

Files

Add a Provider Package

Add a complete first-party @ai-sdk/<provider> package that follows the current provider interfaces, repository conventions, security requirements, and release process.

Read the Current Sources of Truth

Before implementing anything, read:

Add new provider for current package and release requirements
Provider development notes for naming, schemas, and workflow serialization
Provider architecture for the provider abstraction
Secure URL handling when the provider fetches polling, image, audio, video, or other URLs

Use PR #18595 as a recent end-to-end example, but choose the current provider package whose API shape and model types most closely resemble the new provider as the implementation reference.

Third parties can publish provider packages outside this repository. A new first-party @ai-sdk/<provider> package requires prior discussion in an issue. Confirm that agreement exists before implementing the package.

SKILL.md on GitHub