Hi. I've been coding the SaaS application and I wanted to give an update. Studio Pro is 3 modules; Live, API and Workers. I built it with these goals: - The whole system is a standalone SaaS application built on a robust framework (Studio API) that anyone can use to spin-up their own automation service for themselves or others. That means taking into account installation, maintenance, support, payments, subscriptions, and complying with the various licenses by the apps and AI models that the system could use. It will have workflow management and a database to replace services like airtable and make.com and n8n for the supported workflows. Payments and subscriptions are TBD so stay tuned. - It will be open-sourced using the MIT license so you're free to "...use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software...". - LIVE - The Frontend Layer has to be lightweight, deployable as standalone or in an orchestrator like Kubernetes and be brand-able using templates or custom designs. The front-end is a visual interface to the API which provides the main functionality like signup, creating api keys, storing secrets for other APIs, workflow management, etc. n8n would be at this layer. - API - The API layer has to be multi-tenant (organizations), use modern methods like websockets, webhooks, API Keys and HMAC signatures, granular permissions (video:write, service-account:read), multiple user roles (super-admin, admin, member), accept and issue API calls, queue management for worker tasks, fully tested, worker cluster management, deployable standalone or in an orchestrator, etc. The Postgres database is at this layer. - WORKER - The Worker Layer has to be horizontally scalable, fast to spin-up, highly performant, state-of-the-art features and be rooted in keeping costs to a minimum if not zero$. We have brainstormed some really cool ideas like an airtable/make/n8n migration tool and AI workflow builder, ComfyUI API integration for advanced effects, and much more.

This video was made using N8N, Image Creator and Video Creator running on my computer. They directly replace Make.com, Leonardo and JSON2Video respectively.I'll update you when the code is released along with the tutorial video. A little heads-up, if you plan on running Studio on your computer, you'll need 16GB+ of RAM and an Nvidia GPU to do image creation and subtitles. If you don't have that kind of setup, I plan to release a docker image that can build a server on runpod.io, with one click, that will dramatically reduce the cost. Cheers

llama.cpp is a standalone server that can run in Self-Host Studio and replace OpenAI for generating stories and image prompts and any other ChatBot/LLM related tasks. Stay tuned for my results and hopefully, a new release of Studio right behind it. Backstory. I use LMStudio for benchmarking and results comparisons of various LLMs and it's a great tool but it only runs as a desktop app. That's too cumbersome and required some human interaction to set it up so, I won't be adding LMStudio into Self-Host Studio. Cheers

Thanks to @Leroy Garas suggestion, I'm building a workflow that doesn't need Airtable. It always annoyed me that since Studio is self-hosted, it has all the assets. Sending them back and forth to an outside entity seemed wasteful. Here's a couple screenshots of how it will work. Lots of tutorials coming so you might want to enable notifications. I should add that the Build Webpage Form branch is only necessary when using a web page to submit the prompts. n8n has a Form node that could easily handle the prompting but it wouldn't have the process update and approval/redo pages that I plan on making.