@reminix/langchain

Reminix Runtime adapter for LangChain. Serve any LangChain runnable as a REST API.

Ready to go live? Deploy to Reminix Cloud for zero-config hosting, or self-host on your own infrastructure.

Installation

npm install @reminix/langchain @langchain/core

This will also install @reminix/runtime as a dependency.

Quick Start

import { ChatOpenAI } from '@langchain/openai';
import { serveAgent } from '@reminix/langchain';

const llm = new ChatOpenAI({ model: 'gpt-4o' });
serveAgent(llm, { name: 'my-chatbot', port: 8080 });

For more flexibility (e.g., serving multiple agents), use wrapAgent and serve separately:

import { ChatOpenAI } from '@langchain/openai';
import { wrapAgent } from '@reminix/langchain';
import { serve } from '@reminix/runtime';

const llm = new ChatOpenAI({ model: 'gpt-4o' });
const agent = wrapAgent(llm, 'my-chatbot');
serve({ agents: [agent], port: 8080 });

Your agent is now available at:

• POST /agents/my-chatbot/invoke - Execute the agent

API Reference

serveAgent(runnable, options)

Wrap a LangChain runnable and serve it immediately. Combines wrapAgent and serve for single-agent setups.

| Parameter | Type | Default | Description | |-----------|------|---------|-------------| | runnable | Runnable | required | Any LangChain runnable (LLM, chain, agent, etc.) | | options.name | string | "langchain-agent" | Name for the agent (used in URL path) | | options.port | number | 8080 | Port to serve on | | options.hostname | string | "0.0.0.0" | Hostname to bind to |

wrapAgent(runnable, name)

Wrap a LangChain runnable for use with Reminix Runtime. Use this with serve from @reminix/runtime for multi-agent setups.

| Parameter | Type | Default | Description | |-----------|------|---------|-------------| | runnable | Runnable | required | Any LangChain runnable (LLM, chain, agent, etc.) | | name | string | "langchain-agent" | Name for the agent (used in URL path) |

Returns: LangChainAgentAdapter - A Reminix adapter instance

Example with a Chain

import { ChatOpenAI } from '@langchain/openai';
import { ChatPromptTemplate } from '@langchain/core/prompts';
import { wrapAgent } from '@reminix/langchain';
import { serve } from '@reminix/runtime';

// Create a chain
const prompt = ChatPromptTemplate.fromMessages([
  ['system', 'You are a helpful assistant.'],
  ['human', '{input}'],
]);
const llm = new ChatOpenAI({ model: 'gpt-4o' });
const chain = prompt.pipe(llm);

// Wrap and serve
const agent = wrapAgent(chain, 'my-chain');
serve({ agents: [agent], port: 8080 });

Endpoint Input/Output Formats

POST /agents/{name}/invoke

Execute the agent. Input keys are passed directly to the LangChain runnable.

Request:

{
  "input": "Hello, how are you?"
}

Response:

{
  "output": "I'm doing well, thank you for asking!"
}

Streaming

For streaming responses, set stream: true in the request:

{
  "input": "Tell me a story",
  "stream": true
}

The response will be sent as Server-Sent Events (SSE).

Runtime Documentation

For information about the server, endpoints, request/response formats, and more, see the @reminix/runtime package.

Deployment

Ready to go live?

• Deploy to Reminix Cloud - Zero-config cloud hosting
• Self-host - Run on your own infrastructure

Links

• GitHub Repository
• LangChain.js Documentation

License

Apache-2.0