Framework Adapters

Wrap popular AI frameworks for use with the Reminix runtime

Reminix provides adapters that let you wrap agents from popular AI frameworks and serve them using the Reminix runtime. This allows you to:

  • Reuse existing agents — No need to rewrite your Vercel AI or OpenAI-based agents
  • Unified API — All agents expose the same /invoke and /chat endpoints
  • Streaming support — Built-in streaming for all adapters
  • Dashboard integration — Monitor all your agents in one place

Installation

Adapters are provided as a separate package for tree-shaking and optional dependencies:

npm install @reminix/runtime @reminix/adapters
# or
pnpm add @reminix/runtime @reminix/adapters

You only need to install the AI framework(s) you're using:

# For OpenAI
npm install openai

# For Anthropic  
npm install @anthropic-ai/sdk

# For LangChain
npm install @langchain/core @langchain/openai

# For LangGraph
npm install @langchain/langgraph

# For Vercel AI SDK
npm install ai @ai-sdk/openai

Available Adapters

FrameworkAdapter Functions
LangChain.jsfromLangChainAgent, fromAgentExecutor, fromLangChainChatModel
LangGraph.jsfromCompiledGraph
Vercel AI SDKfromToolLoopAgent, fromGenerateText
OpenAI SDKfromOpenAI
Anthropic SDKfromAnthropic

Import Styles

Subpath imports (recommended for tree-shaking and explicit dependencies):

import { fromOpenAI } from '@reminix/adapters/openai';
import { fromAnthropic } from '@reminix/adapters/anthropic';
import { fromAgentExecutor } from '@reminix/adapters/langchain';
import { fromCompiledGraph } from '@reminix/adapters/langgraph';
import { fromGenerateText } from '@reminix/adapters/vercel-ai';

Main entry import (access all adapters via a single import):

import { fromOpenAI, fromAnthropic, fromGenerateText } from '@reminix/adapters';

Quick Example

import OpenAI from 'openai';
import { fromOpenAI } from '@reminix/adapters/openai';
import { serve } from '@reminix/runtime';

const client = new OpenAI();
const agent = fromOpenAI(client, {
  name: 'gpt4-agent',
  model: 'gpt-4o',
});

serve(agent);

That's it! Your OpenAI-powered agent now has:

  • POST /agent/gpt4-agent/invoke — Single-turn requests
  • POST /agent/gpt4-agent/chat — Multi-turn conversations
  • Streaming support with stream: true
  • Health checks and discovery endpoints

On this page