ai-lcr
Integrations

Overview

lcr("...") returns a standard AI SDK model, so anything that accepts one works unchanged — the Vercel AI SDK itself, Mastra agents, and any framework built on the AI SDK.

The reason ai-lcr integrates so widely is one property: lcr("...") returns a standard AI SDK LanguageModel. Anything that takes an AI SDK model takes a routed one — you change the model argument and nothing else.

New here? Migrate from plain AI SDK shows the core one-line swap. The integrations below all build on it.

Pick your framework

  • Vercel AI SDK — generateText, streamText, generateObject, tools, and the Agent loop.
  • Mastra — drop lcr(...) into a Mastra agent's model field.

Other frameworks

Any library that accepts an AI SDK LanguageModel works the same way — pass lcr("...") where it expects a model. Frameworks that use their own model classes (e.g. LangChain's chat models) or run outside JavaScript (Python tools, visual flow builders like n8n) don't consume an AI SDK model directly; bridging those needs an OpenAI-compatible proxy in front of ai-lcr, which is on the roadmap rather than shipping today.

Runware

Runware — cheap image generation via a single synchronous call. Image-only — no video. A media provider, wired with createRunwareMediaAdapter.

Vercel AI SDK

lcr("...") drops into every Vercel AI SDK call — generateText, streamText, generateObject, tools, and the Agent loop — unchanged.

On this page

Pick your frameworkOther frameworks