Integrations

LangChain

Use DEVUP AI models with LangChain's ChatOpenAI class for seamless Python AI workflows.

Because DEVUP AI exposes a fully OpenAI-compatible API, you can use it as a drop-in replacement in any LangChain pipeline. Point your openai_api_base to our gateway and access 142+ open-source models with zero code changes to your existing chains.

Prerequisites

A DEVUP AI account with a valid API key (Dashboard → API Keys).
Python 3.9 or later.
The langchain-openai package installed.

Installation

bash

pip install langchain-openai

Basic Usage

Pass your DEVUP API key and gateway URL to ChatOpenAI. All 142+ models available on the platform are accessible via the model_name parameter.

python

from langchain_openai import ChatOpenAI

# Initialize the DEVUP AI endpoint
chat = ChatOpenAI(
    openai_api_key="your-devup-api-key",
    openai_api_base="https://api.devupai.com/v1",
    model_name="deepseek-chat"  # or any of our 142+ models
)

response = chat.invoke("Hello, DEVUP AI!")
print(response.content)

Using Environment Variables

For production deployments, store your API key in an environment variable instead of hardcoding it:

bash

export DEVUP_API_KEY="sk-devup-your_key_here"

python

import os
from langchain_openai import ChatOpenAI

chat = ChatOpenAI(
    openai_api_key=os.environ["DEVUP_API_KEY"],
    openai_api_base="https://api.devupai.com/v1",
    model_name="deepseek-chat",
)

Streaming Responses

Enable streaming to receive tokens as they are generated. This is ideal for real-time UIs and long-form content generation:

python

from langchain_openai import ChatOpenAI

chat = ChatOpenAI(
    openai_api_key="your-devup-api-key",
    openai_api_base="https://api.devupai.com/v1",
    model_name="deepseek-chat",
    streaming=True,
)

for chunk in chat.stream("Explain quantum computing in 3 sentences."):
    print(chunk.content, end="", flush=True)

Building Chains with LCEL

DEVUP AI works seamlessly with LangChain Expression Language (LCEL). Compose prompts, models, and output parsers into production pipelines:

python

from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate

chat = ChatOpenAI(
    openai_api_key="your-devup-api-key",
    openai_api_base="https://api.devupai.com/v1",
    model_name="deepseek-chat",
)

prompt = ChatPromptTemplate.from_messages([
    ("system", "You are a senior {language} developer."),
    ("human", "{question}"),
])

chain = prompt | chat

response = chain.invoke({
    "language": "Python",
    "question": "What is the best way to handle async in Python 3.12?"
})

print(response.content)

Configuration Options

Fine-tune model behavior using standard OpenAI parameters:

python

chat = ChatOpenAI(
    openai_api_key="your-devup-api-key",
    openai_api_base="https://api.devupai.com/v1",
    model_name="deepseek-chat",
    temperature=0.2,         # Lower = more deterministic
    max_tokens=2048,         # Max completion length
    request_timeout=30,      # Timeout in seconds
)

Parameter	Type	Description
openai_api_base	str	Always `https://api.devupai.com/v1`
openai_api_key	str	Your DEVUP AI API key (prefixed with `sk-devup-`).
model_name	str	Any supported DevUp model alias or full model ID.
temperature	float	Sampling temperature (0.0–2.0). Default: 0.7.
max_tokens	int	Maximum number of completion tokens to generate.
streaming	bool	Enable token-by-token streaming. Default: False.

Notes

DEVUP AI is fully OpenAI-compatible — any LangChain component that uses ChatOpenAI or OpenAIEmbeddings works out of the box.
Tool calling, structured outputs, and function calling are supported on models that natively support them.
All requests are metered and billed in Algerian Dinar (DZD) via your account balance.

If you are migrating from OpenAI, the only change required is setting openai_api_base to https://api.devupai.com/v1 and swapping your API key. All existing chains, agents, and tools remain fully compatible.

AutoGen

Authentication