​
Introduction

Julep’s Open Responses is a self-hosted, open-source implementation of OpenAI’s Responses API that works with any LLM backend. It provides a lightweight interface for generating content with Large Language Models (LLMs) without needing to create persistent agents or sessions.

To try it out, just run npx -y open-responses init (or uvx) and that’s it! :)

​
What is Open Responses?

Julep’s Open Responses lets you run your own server that is compatible with OpenAI’s Responses API, while giving you the freedom to use alternative models like:

  • Anthropic’s Claude
  • Alibaba’s Qwen
  • Deepseek R1
  • and many others …

It’s essentially a drop-in replacement that you control, with a permissive Apache-2.0 license. As an early release, we welcome your feedback and contributions to help improve it.

Open Responses API Overview

​
Why Open Responses?

  • Model Flexibility: Use any LLM backend without vendor lock-in, including local model deployment
  • Self-hosted & Private: Maintain full control over your deployment on your own infrastructure (cloud or on-premise)
  • Drop-in Compatibility: Seamlessly integrates with the official Agents SDK by simply pointing to your self-hosted URL
  • Easy Deployment: Quick setup via docker-compose or our CLI with minimal configuration
  • Built-in Tools: Automatic execution of tool calls (like web_search) using open & pluggable alternatives
  • The Open Responses API requires self-hosting. See the installation guide below.
  • Being in Alpha, the API is subject to change. Check back frequently for updates.
  • For more context, see the OpenAI Responses API documentation.

​
Local Installation

This section will guide you through the steps to set up the Julep’s Open Responses API.

​
Prerequisites

Install Docker

​
Installation

The Julep’s Open Responses API is a fully microservice-based architecture. It is fully dockerized and can be easily deployed on any infrastructure that supports Docker. There are two ways to install the API:

​
Docker Installation

1

Create a directory for the project

mkdir julep-responses-api
2

Navigate to the project directory

cd julep-responses-api
3

Download and edit the environment variables

wget https://u.julep.ai/responses-env.example -O .env

Edit the .env file with your own values.

4

Download the Docker Compose file

wget https://u.julep.ai/responses-compose.yaml -O docker-compose.yml

Download the file to the current directory with the name docker-compose.yml. This is the file that will be used to run the Docker containers.

5

Run the Docker containers

docker compose up --watch

This will start the containers in watch mode.

6

Verify that the containers are running

docker ps

​
CLI Installation

The CLI is a lightweight alternative to Docker for those who prefer not to use Docker directly.

Internally, it uses Docker to run the containers.

1

Install the CLI

You can install the CLI using several package managers:

# Using npx directly
npx open-responses

# Or install globally
npm install -g open-responses
2

Setup the Environment Variables

npx open-responses setup

Before using any commands, you must run the setup command

3

Run the CLI

npx open-responses start

This will start the API in watch mode

To learn more about the CLI one can use the checkout the CLI Documentation.

​
Quickstart Example

With the OpenAI SDK initialized, you can now use the Responses API to generate content.

API Key Configuration

  • RESPONSE_API_KEY is the API key that you set in the .env file.

Model Selection

  • While using models other than OpenAI, one might need to add the provider/ prefix to the model name.
  • For supported providers, see the LiteLLM Providers documentation.

Environment Setup

  • Add the relevant provider keys to the .env file to use their respective models.

​
1. Install the OpenAI SDK

pip install openai

​
2. Initialize the OpenAI client

from openai import OpenAI
openai_client = OpenAI(base_url="http://localhost:8080/", api_key="RESPONSE_API_KEY")

​
3. Generate a response

import os
from openai import OpenAI

openai_client = OpenAI(base_url="http://localhost:8080/", api_key=os.getenv("RESPONSE_API_KEY"))

response = openai_client.responses.create(
    model="gpt-4o-mini",
    input="How many people live in the world?"
)
print("Generated response:", response.output[0].content[0].text)

​
Next Steps

You’ve got Open Responses running – here’s what to explore next: