@xynqai/langchain

LangChain.js / LangGraph integration for XYNQ — decentralized, OpenAI-compatible inference.

npm install @xynqai/langchain @langchain/openai @langchain/core

Quick start

import { ChatXYNQ } from "@xynqai/langchain";

process.env.XYNQ_API_KEY = "xynq_sk_...";

const llm = ChatXYNQ.smart({ temperature: 0.7 });
const reply = await llm.invoke("Explain backpropagation in two sentences.");
console.log(reply.content);

Get a key at xynq.ai → Account (1M free tokens/month).

API

ChatXYNQ

Drop-in ChatOpenAI subclass with XYNQ defaults:

| Default | Value | Override | | --- | --- | --- | | Base URL | https://xynq.ai/v1 | XYNQ_BASE_URL or baseURL | | API key | XYNQ_API_KEY | apiKey | | Model | xynq-1 | model |

import { ChatXYNQ, FAST_MODEL, SMART_MODEL } from "@xynqai/langchain";

const fast = ChatXYNQ.fast();
const smart = ChatXYNQ.smart();
const custom = new ChatXYNQ({ model: "xynq-qwen-3.5-27b", temperature: 0 });

Streaming:

const stream = await ChatXYNQ.smart().stream("Count to five.");
for await (const chunk of stream) {
  process.stdout.write(chunk.content ?? "");
}

LangGraph

import { createReactAgent } from "@langchain/langgraph/prebuilt";
import { ChatXYNQ } from "@xynqai/langchain";

const agent = createReactAgent({
  llm: ChatXYNQ.smart(),
  tools: [...],
});

Models

| Helper | Model id | | --- | --- | | ChatXYNQ.fast() | xynq-1 | | ChatXYNQ.smart() | xynq-glm-5.2 | | custom | xynq-qwen-3.5-27b, xynq-supergemma-4-26b |

Python

Use langchain-xynq on PyPI.

Related

• XYNQ API reference
• @xynqai/plugin-eliza
• LangGraph examples

License

MIT