Full-Stack Generative Server with Agent and Studio

Run your private generative cloud with your own API keys.

Full-control over access and storage.

One-time purchase license, downnload minor updates free.

We also provide support regarding models and configurations.

Join our discord for support.

⚠️ Currently in test-preview. Message us for access.

Visit our GA web-app at https://app.samsar.one

Join our Discord community for early access and updates.

---

🧠 Overview

This repository provides containers for running the Generative Agent + Studio pipeline in professional or enterprise server environments.

Download and run the Docker images with your own environment configurations.

---

🚀 Installation

1. Ensure Docker is installed on your system.
2. Acquire the API keys for the models you wish to use.
3. Run the containers using:

docker compose up

---

🔐 Environment Variables Setup

Configure the following environment variables before starting:

🔹 OpenAI Keys

• Required for agent inference.
• The studio lets you choose between different OpenAI models.

🔹 Fal AI Keys

• Most media models (image, video, etc.) use Fal AI for inference.
• Add your Fal API key to the environment.

🔹 Azure Keys

• Required for using models like Sora.
• Add both the Azure endpoint URL and the API key.

Example: Access Sora via Azure Foundry Sora

For other models, get the corresponding connection strings and endpoints from Azure.

🔹 RunwayML Keys

• Needed if you plan to use RunwayML as video generation model.

🔹 Replicate Keys

• Needed if you plan to use Audiocraft as the music generation model.

---

📦 Storage Configuration

• Add AWS Access Keys for storing media (videos, frame images, etc.).
• Also configure a CloudFront URL for accessing stored assets.

---

Let me know if you want this version saved as a file or want it tailored to a specific use case (e.g. enterprise onboarding, local dev setup, etc.).