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.

Hello. I've been doing a lot of n8n workflows for myself and others and I have run into two frustrating issues; auto-saving workflows I'm working on and auto-backups of saved workflows. I'm a software developer and site reliability engineer so, naturally, I wrote some apps to handle that. Workflow Auto-Save is a Chrome browser extension. It detects when the Save button is activated on an n8n workflow and based on a time interval you set, it clicks the Save button, automatically. See screenshot for a view of the app. Workflow Auto-Backup is a shell script that detects changes to all your saved workflows. When it detects a change, it creates a backup of the workflow with a timestamp added to the name. They're both available at https://github.com/kickin6/studio-utils. The Workflow Auto-Save Browser Extension has been submitted to the Chrome Web Store for review and hopefully, approval. 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

I'm making Template Repository public today. It is the framework that all Self-Host Studio apps and modules will be based on and in. It can be a top-level repository or a sub-module. I'll use it for both. I'll put together a demo later this week. Some features: - Implements API best practices. - Dynamic input payload validation. - JSON to JSONSCHEMA conversion script to convert Make.com mappable parameters into a format that makes testing and validation automatic. - Dynamic unit and integration testing. - Fully modular for plug-in support. - Automatic API documentation through /docs URL. - Built-in scalable queueing system for apps that don't have one like FFMPEG. - Webhook inbound and outbound support. - Can run standalone, docker and in kubernetes environments. - Lightweight. Docker images are about 500MB. - Easy to configure using environment variables. It's available at https://github.com/kickin6/selfhost-template It's a Github template so the easiest way to get it Use this template button (pic below). 🚗 Roadmap - Migrate Self-Host Image and Self-Host Video to this platform. - Implement queue management UI using Flower.