frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

CorridorKey 2.0

https://corridorkey.com/
95•ravenical•1h ago

Comments

lambfoo•52m ago
It is actually free for people other than organizations with more than US $20 million in annual revenue.

Source: https://github.com/nikopueringer/CorridorKey2.0

DetroitThrow•42m ago
Wow, that's amazing to try in casual visual editing workflows. Very great results as well for something free.
teach•40m ago
I also really enjoyed the "making of" video: https://youtu.be/svrdRAoIlxM?si=cG1gSRyWdGg8RnEM

Niko goes into some more detail about the licensing choices they've made and why towards the end of the video as well.

[Edit: note that the above video is "content" so if you're not used to the style it might be offputting but I'm into it]

It's been so cool to watch this develop

CitrusFruits•18m ago
Huge fan of Corridor. For those that don't have twenty minutes, one of the big takeaways is that a pretty high level VFX Supervisor critiques the keys at the end and basically has only nits that you need to zoom in on or have a trained eye to catch. Very legitimate feedback for a top tier professional context, but this can get you really far otherwise.
ipsum2•13m ago
The issues aren't nits, they're really obvious, even after YouTube compression.
emkoemko•9m ago
did i watch the same video? it was no where near professional level, he even said, as a first pass pre vis for test shots it would be great
xfeeefeee•33m ago
I'd love to use this but sadly my local GPU is just crap. I know everyone hates hosted offerings but I'd be a happy customer; I mostly use topaz's hosted stuff too for this reason
ShakataGaNai•31m ago
A beta price of $50 or full price of $100 is really a steal. If you do ANY sort of video production work, this is an amazing tool. The more casual of a green screen person you are, the more impressive this is.

This is like Magnet Mask for Final Cut, but probably better. I haven't compared/contrasted the two yet, but I have faith that Corridor has knocked this one out of the park from the demos I have seen.

jarjoura•31m ago
Just to be clear, this isn't generative AI in the same way something like Nano Banana or Seedance is. You don't prompt anything into existence. However, this is using all of the latest SOTA techniques that the slop generative models also depend on.

I think it is an amazing example of how labs should take all of their fundamental research and turn into products entirely designed, limited and scoped only to empower artists instead of obsessed with trying to bypass them.

DrewADesign•21m ago
> labs should take all of their fundamental research and turn into products entirely designed, limited and scoped only to empower artists instead of obsessed with trying to bypass them.

The software companies aren’t going to start doing this because the media companies aren’t interested in doing that. Tools like Nuke that already do this are expensive tools for people that the tech industry is trying to make obsolete. The media companies want to sell anything that will sell, and spend as little as possible in the process.

I’m pretty sure this will have to come from skilled and experienced artists doing amazing things with technology despite the tech industry sabotaging their careers, which in turn develops demand for actual art rather than bulk-created ‘content’. Sadly, I think we’re years away from that. I graduated in this field in 2024 and my career was dead before it started. Most people in this business with the hard-won skills to do this work that spent years developing their taste, storytelling sophistication, and craft are in the process of being systematically fucked over by the tech industry. Sad.

imaginationra•14m ago
Already obsolete- who keys anymore besides youtube slop peddlers?

Just go full cg or full ai gen.

rtkwe•10m ago
Movies are absolutely full of background replacement and other things that require keying.
imaginationra•1m ago
So its for mainstream film studios? Thats tired status quo stuff.
ilaksh•4m ago
I think there is still a little bit of room for green screen. At least maybe until Seedance 3.0 in a few months.

But it is interesting how video generation turned the corner from several months ago to now. I think the next AI project they do probably has to be with a generative tool even though they may be very hesitant to do that because they don't want to emphasize AI models that threaten their own jobs.

But I remember when they did the Project Greenlight thing which resulted in a massive list of desired projects from subscribers that they could not possibly make a dent in.

I think in the next several months they may be able to revisit that or at least some of the projects that had to be shelved due to lack of time. To make that possible they would leverage the latest generative AI models and workflows, possibly a few that haven't been released yet. But we are rapidly approaching the point where the models plus tooling is good enough to replace 95% of the technical work, and it's down to creativity. I think it's only a couple of years before that progress bar really maxed out.

snitty•13m ago
On the first image the arrows toggle between green screen and rotoscoped. On the second set of images, the arrows toggle between the images.
lrvick•13m ago
I hope they open source it before someone else decompiles it and does it for them. Just charge for binaries that most will pay for anyway. Then things like Linux support could be handled by the community.
scblock•9m ago
https://github.com/nikopueringer/CorridorKey2.0
ShowalkKama•8m ago
it's already open source, only the plugins are paid.
LeoWattenberg•4m ago
It is source-available, not open source.
CiaranMcNulty•8m ago
The core tool is CC BY-NC-SA 4.0

They’re charging for the bundled plugins for various editing suites

LeoWattenberg•4m ago
It is not, there's additional restrictions beyond the CC license.
paaloeye•3m ago
I always thought CorridorKey was free and open-source

Working with academic communities to build scientific understanding

https://openai.com/index/building-scientific-understanding-with-academic-communities/
163•jonluca•1h ago•42 comments

CorridorKey 2.0

https://corridorkey.com/
97•ravenical•1h ago•24 comments

Type Server Protocol

https://github.com/microsoft/pylance-release/blob/main/docs/tsp/type-server-protocol.md
169•adamnemecek•4h ago•125 comments

Sebastian Raschka's from-scratch video course on building reasoning models

https://magazine.sebastianraschka.com/p/building-reasoning-models
502•sgiannopoulos•7h ago•76 comments

CSS-Tricks has a new owner

https://css-tricks.com/css-tricks-has-a-new-owner/
344•edent•7h ago•79 comments

Cross-request KV cache exposure in the GLM-5.3 Workers AI model (155 responses)

https://blog.cloudflare.com/kv-cache-exposure-remediation/
10•wondersky•14m ago•1 comments

Oldest GUI application that still runs on Windows 11

https://blog.enlyth.com/posts/oldest-windows-gui-app-win-11/
223•enlyth•4h ago•131 comments

Nietzschian Debugger

https://github.com/DilawarShafiq/nietzschian-debugger
111•dilawargopang•5h ago•37 comments

A Printer Printing onto Paper

https://twitter.com/macrofying/status/2112904628509528219
9•MrBuddyCasino•3d ago•1 comments

GPT-6.1 Terra

https://openai.com/index/introducing-gpt-6-1-terra/
445•meetpateltech•4h ago•283 comments

Show HN: Hob - Open-source, on device alternative to Claude Cowork and ChatGPT

https://github.com/FrigadeHQ/hob
76•pancomplex•6h ago•21 comments

Anthropic S-1

https://www.sec.gov/Archives/edgar/data/2130357/000162828026067168/anthropicpbc-sx1.htm
1407•SiNTEx•9h ago•1375 comments

Native Prometheus Integration with Nginx

https://categorica.io/blog/2026.10.16_nginx-stats-prometheus/
20•foxhill•3d ago•5 comments

A Man from the Future: Why Johnny Can't Stop Abstracting

http://muratbuffalo.blogspot.com/2026/10/the-man-from-future-why-johnny-cant.html
91•aray07•5h ago•45 comments

Diamond Transistor

https://www.df.com/diamond-transistor
108•Gravityloss•3d ago•59 comments

SIMD-accelerated computer vision on a $2 microcontroller

https://shraiwi.com/blog/simd-fast-esp32s3/
157•zdw•1d ago•52 comments

Microsoft offers mass buyouts for roughly a quarter of its workforce

https://www.seattletimes.com/business/microsoft/microsoft-offers-mass-buyouts-for-roughly-a-quart...
661•doctorzook•7h ago•555 comments

Blaise Pascal invented fixed route public transport (2024)

https://humantransit.org/2024/12/one-of-the-greatest-inventions-in-the-history-of-human-transport...
251•skadamat•4d ago•133 comments

Birdwatching rewires the brain and slows cognitive decline

https://www.bbc.com/future/article/20261019-birdwatching-rewires-the-brain-and-slows-cognitive-de...
488•cisc•3d ago•191 comments

Pharo 14

https://pharo.org/news/2026-10-26-Pharo14Release.html
228•xkriva11•11h ago•57 comments

In Leaked Audio, Flock CEO Tells Cops Backlash to the Company Isn't Real

https://www.404media.co/in-leaked-audio-flock-ceo-tells-cops-backlash-to-the-company-isnt-real/
233•Frieren•5h ago•141 comments

The EPA Is Quietly Killing Its Web Archive

https://www.motherjones.com/environment/2026/10/epa-online-archive-deletion-zeldin-camp-lejeune-t...
187•robtherobber•3h ago•30 comments

Who killed the EV cert? A browser whodunit for Halloween

https://qwac.com/blog/who-killed-the-ev-cert
15•roddylindsay•3d ago•27 comments

US designates helium as a critical material

https://apnews.com/article/energy-critical-materials-minerals-helium-43da160adc1bad389fba0b9ffbd1...
151•rbanffy•4h ago•87 comments

Magic state cultivation on a superconducting quantum processor

https://arxiv.org/abs/2512.13908
7•kheal•3h ago•0 comments

Show HN: Glider 4.0 Unofficial Web Edition (Mac game port)

https://dbohdan.github.io/glider/
29•networked•23h ago•5 comments

The Exponential Time Hypothesis Is False [pdf]

https://github.com/openai/math/blob/main/preprints/The-Exponential-Time-Hypothesis-is-False-Octob...
12•vok•56m ago•0 comments

Switzerland. Block by block. At 1:1 scale

https://blocks.manifold3d.ch/
450•theanonymousone•14h ago•212 comments

Concerning behavior from a popular open-weight AI model

https://cohere.com/blog/concerning-behaviour-popular-open-weight-ai-model
238•matthieu_bl•6h ago•96 comments

Fixing Tool-Call Accuracy in vLLM

https://www.ubicloud.com/blog/fixing-tool-call-accuracy-in-vllm
143•coffeemug•9h ago•64 comments