Skip to main content

LangChain

Use ChatOpenAI with Djelia's base URL. The language pair goes in extra_body, the same as with the OpenAI SDK.

import os
from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    model="banjugu-1",
    api_key=os.environ["DJELIA_API_KEY"],
    base_url="https://api.djelia.cloud/openai/v1",
    extra_body={"djelia": {"source_language": "fra_Latn", "target_language": "bam_Latn"}},
)

print(llm.invoke("Bonjour, comment allez-vous ?").content)

In a chain​

Because the language pair is set on the model rather than per call, a Djelia translator composes like any other runnable:

from langchain_core.prompts import ChatPromptTemplate

to_bambara = ChatOpenAI(
    model="banjugu-1",
    api_key=os.environ["DJELIA_API_KEY"],
    base_url="https://api.djelia.cloud/openai/v1",
    extra_body={"djelia": {"source_language": "eng_Latn", "target_language": "bam_Latn"}},
)

chain = ChatPromptTemplate.from_template("{text}") | to_bambara
chain.invoke({"text": "Welcome to the clinic."})
caution

Djelia's translation model has no tool use, no system-prompt steering and no conversational memory. It translates the last user message. Agent abstractions that expect tool calling will not work against it.