Hello Everyone!

So I wanted to introduce everyone to an AGPL-3.0 OSS, free forever, AI Video Studio that I've been working on for the past few months called "Vivijure." Vivijure is a web based modular frontend that allows you to collaborate with AI to plan your videos, create images of your cast members and generate all the images for a full LoRA training set, and render your videos either on your own iron (LTX Video) or rent by the second serverless workers (Wan 2.2). It also has integration with several Cloud i2v providers through the interface and everything gets planned in a standard storyboard.json format. I've integrated Real-ESRGAN for video upscaling, MuseTalk for lip syncing your characters to their spoken lines, audio upscaling for speech (GPU based) and music (CPU based). The really cool thing that I've implemented for the design is it's not wasteful when it comes to your GPU spend, anything that can be done by a CPU only, Vivijure routes the job to one of its CPU worker pods instead of you wasting your money or electricity running a CPU-centric job on a GPU.

There's also a Discord Bot I've designed to go with it called Slate that allows you and your friends to plan, research, and render videos together on your own Discord Server. Vivijure supports create title and credit cards, subtitles, narration, music generation, and timing your shots to music that you upload yourself.

Right now the front end control panel runs as a Cloudflare Worker, along with the modules, and it connects to R2 and D1 for data processing and storage. Anybody can write a module as long as it follows the contract set out (and extensively documented). I'm working on a cloud agnostic version of the control panel as part of the next major release, but for now, the front end is Cloudflare powered and for the basic set of modules you can host it for free. To integrate the video upscaling, MuseTalk, and speech upscaling modules you do have to upgrade to Cloudflare's Workers Paid plan, which is $5/month.

There's a limited demo that I've setup if you would like to see how the interface looks and feels at https://demo.vivijure.com. You can find more info, including a "ask me anything about Vivijure" widget at https://vivijure.com, and as always, the source code is available at https://github.com/skyphusion-labs/vivijure.

I really look forward to people checking this out and trying it out, I really enjoyed working on Vivijure and it's my gift to the world - free yourself from SaaS AI Video Studio providers, regain your privacy and control over your own data, and most of all, be creative without being gated by entities that want to limit your vision. There's an easy to use installation script that comes with Vivijure, it should take you no more than an hour to be fully operational with it.

If anyone has any feedback, questions, or comments, please let me know. I'm very excited to show this to the world and I hope someone could gain good use from this software.

Thanks,

Skyphusion (Conrad Rockenhaus)

Show HN: ProtoLink – A Python runtime for distributed A2A agents

Browser over SSH

How the most impactful AI startups will be built in emerging markets

Archangel: CIA's SUPERSONIC A-12 RECONNAISSANCE AIRCRAFT [pdf]

SociaLLM Engineering: On Manipulating AI Agents and what we can do about it

Toast IDE Gets LSP Goto and Autocomplete

Cairn, an AI agent with a $50 budget, an email address, and a constitution

TrulyFreeOCR – Permissive-license OCR to PDF, self-contained, MRC compression

Show HN: Runscribe – Turn a terminal session into a re-runnable runbook, offline

I'm Making Strandfall, a Solarpunk Orienteering Larp

[Show HN] Latent-free ternary LLM training

A Farewell to ARPs: IPv4 Service on IPv6-Only Networks

Is Vibecoding Viable for Business?

Deaths in the UK outnumber births for first time in 50 years

MCP Spec Check

SMS Pumping Fraud

Apple accuses OpenAI of stealing its core tech secrets

Make LLM-powered apps Smarter and Cheaper, Reliable, Auditable changing 1 LOC

Guide to Response Time Monitoring

Minlz v1.2: Search compressed data with without decompression

"Houston, we have 1 problem(s)."

Anduril CEO Warns AI and Defense Valuations Are Becoming Dangerously Overvalued

Musk's Starlink Socks Customers with $1500 'High Demand' Surcharge

Open source AI matters more than ever, according to Hugging Face's Clem Delangue

Show HN: A Sims-style house builder in the browser (Three.js, no back end)

Show HN: OSS Web Based AI Video Studio (Vivijure)

I Stopped Destructuring Everything

Show HN: Costbase – See what each of your AI apps costs, without a proxy

The Emperor's New Xerox

Show HN: Hypothesis Prism, a graph representation of a paper's hypothesis

Ask HN: Is GPT-5.5 being nerfed?