Other OpenAI-compatible tools

Maple Proxy works with any tool that lets you set a custom OpenAI base URL and uses the chat completions, embeddings or models endpoints. Look for settings named base URL, API base, host or custom OpenAI provider.

Most tools need the same three values:

Some tools want the host without /v1; if requests fail with 404, try the other form.

LangChain (Python)

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    base_url="http://127.0.0.1:8080/v1",
    api_key="your-maple-api-key",
    model="gpt-oss-120b",
)

LlamaIndex

Use the OpenAILike class, which accepts model names that OpenAI doesn’t know:

from llama_index.llms.openai_like import OpenAILike

llm = OpenAILike(
    api_base="http://127.0.0.1:8080/v1",
    api_key="your-maple-api-key",
    model="gpt-oss-120b",
    is_chat_model=True,
)

Open Interpreter

Current Open Interpreter reads custom providers from ~/.openinterpreter/config.toml. Add Maple Proxy as a Chat Completions provider:

model_provider = "maple"
model = "gpt-oss-120b"

[model_providers.maple]
name = "Maple Proxy"
base_url = "http://127.0.0.1:8080/v1"
env_key = "MAPLE_API_KEY"
wire_api = "chat"

Then set MAPLE_API_KEY in your shell and run interpreter.

These examples follow each library’s current documentation but haven’t been run against the proxy. If a setting has moved, check that library’s docs for its custom OpenAI base URL option.

What won’t work

The proxy forwards only chat completions, embeddings and the model list. Tools that need OpenAI’s Responses API, Assistants, audio, images or file uploads get 404 from the proxy.