langchain-empiriolabs

This package contains the LangChain integration for EmpirioLabs AI.

EmpirioLabs AI is specialized AI model hosting for open, proprietary, and custom stacks. Its chat models include Qwen3, DeepSeek V4, GLM-5.1, Kimi K2.7 Code and MiniMax M3; browse the full model catalog at docs.empiriolabs.ai.

Installation

pip install -U langchain-empiriolabs

Set your API key as an environment variable:

export EMPIRIOLABS_API_KEY="your-api-key"

You can create a key from the EmpirioLabs AI dashboard.

Chat models

ChatEmpirioLabs wraps the EmpirioLabs chat completions API.

from langchain_empiriolabs import ChatEmpirioLabs

llm = ChatEmpirioLabs(model="qwen3-7-plus")

response = llm.invoke("Explain backpropagation in one paragraph.")
print(response.content)

Streaming:

for chunk in llm.stream("Write a haiku about the sea."):
    print(chunk.content, end="", flush=True)

Tool calling:

from pydantic import BaseModel, Field

class GetWeather(BaseModel):
    """Get the current weather in a given location."""

    location: str = Field(..., description="The city and state, e.g. San Francisco, CA")

llm_with_tools = llm.bind_tools([GetWeather])
print(llm_with_tools.invoke("What is the weather in SF?").tool_calls)

Available models

Any model slug from the EmpirioLabs AI catalog works. Some popular chat slugs:

Reasoning models place their chain-of-thought in response.additional_kwargs["reasoning_content"].

The live list is always available from GET https://api.empiriolabs.ai/v1/models.

Embeddings

EmpirioLabsEmbeddings wraps the EmpirioLabs embeddings API.

from langchain_empiriolabs import EmpirioLabsEmbeddings

embeddings = EmpirioLabsEmbeddings(model="text-embedding-v4")

vector = embeddings.embed_query("Hello, world!")
print(len(vector))

A request takes up to 2,048 texts; embed_documents sends 1,000 per request by default (chunk_size).

Configuration

License

MIT