itch.io

> I noticed yesterday that when I kept clicking the icon on the desktop, you get at least two opening apps.

I'll try putting a cap on the number of open apps.

> I was wondering if you might be thinking about us being able to upload an image that can be lip synced and saved as a video.

I was actually thinking of this! I want to build this for myself so I can make talking head videos to show off VCP on social media.

Good to hear the errors are gone.

> Oh, one more thing but I'm guessing it's not on your end, but the Omnivoice [sigh] has never worked for me.

Some of the Omnivoice tags are really hit or miss. They also don't always work unless they're in a place that the model expects it. For instance if the model was trained on audio that only has sighing at the end of the clip, you're far more likely to get a working sigh if you add the tag at the end of a sentence. 

If you want the most expressive speech try DramaBox in the design tab. It needs more work to get right but you can get some very...dramatic speech out of it. 

> When you have it crash, say five times, you have five in task manager and the system tray.

I think I've isolated this so the next update should fix it. Can you either DM me your log files on Discord (@orville) or send them to the support email? I want to see what's been causing all your crashes. To get logs open settings and press the "Logs" button. 

> Also I just noticed that Kokoro is not under the switch model drop down box on the main interface.

This is because the app is structured by feature (Clone/Design/TTS) but not every model supports that feature. Qwen3 can clone your voice, create voices through text descriptions, and has built-in voices so it appears in all 3 tabs but since Kokoro only has built-in voices, it only appears in the TTS tab. 

Is this potentially confusing? Yes, yes it is, and I'm trying to figure out a way to clear this up. Open to ideas.

> If this is supossed to be offline, how come the app won't start unless I'm connected to the internet?

After entering the license key if your hardware doesn't change then it should launch while offline. This might sound silly but how long did you wait on the spinning wheel? Sometimes the app can take up to a minute to launch. 

> Is there any way we can get the ability to right click cut and paste in the text box?

Will add this to the next update!

You're right and I'll keep looking for smaller models. My target is always 8GB of VRAM :)

Hello again 👋

I just released FX Creator Pro as a small freebie for people who bought my other 2 apps, Voice Creator Pro or Song Creator Pro. 

FX Creator lets you describe any sound in plain text (Foley, nature and weather, creatures, machinery, magic and sci-fi, and more) and generate high-quality audio in seconds. It works really well for ambiance up to 30 seconds long (chirping birds, rolling thunder, etc.) but can also make shorter clips for things like explosions or gun shots. 

It's free for previous owners and $10 for everyone else.

Yeah, unfortunately quantization has to be disabled for LoRAs. I still like the idea so I will add it to Song Creator :)

I see. I think the first link you shared is an entire SFT model. The second does look like a LORA trained on psytrance. Let me look into adding support for this, I like the idea of easy LORA import. 

Do you mean you want to train your own LORA or use existing ones?

No problem and sorry once again

Hey Jerome see my last reply here.  In short, it's currently gated to 16GB cards.

Hey Jerome. Do you have at least 16GB of VRAM? The models may work with 12GB but inference would likely be painfully slow so I decided to make them available only to larger cards. I should have made this clearer in the post so my apologies.

Having said that, I could make the models available on 12GB cards if you're keen on trying them out and are willing to put up with potential slowness. 

Good to know, I'll add it to my list

Awesome!  My record is generating an 11 hour long audiobook without major issues. 

I did briefly experiment with a higher quality 48khz output for qwen3 but it introduced pronunciation errors  so I removed it. Might take a look again if you think it's worth it. 

No problem. Let me know if it works. I'll work on making it more obvious that you should use projects for long-form content.

What model are you using? Chatterbox tends to degrade over longer speech.

Also, are you using the Lab or Projects interface to generate speech? For text that's longer than a few sentences you should create a project so VCP can properly chunk the text.

The one that says version 1.7.5 is the latest (for now). You don't need both, I left the second in case the latest version doesn't work. 

If you get lifetime you'll get updates as long as I release them (or get hit by a bus). 

Itch should notify you by email when there's a new update. You should also get emails when I make a new dev post. If you join Discord you can get notified there too.

Ae you still seeing this problem in the latest version?

Hey Thomas, so sorry about the late reply but Itch flagged your post as spam for some reason :(

To control speed you can tweak advanced settings (the icon with the sliders) for the models.

Chatterbox: Increase "Exaggeration"
OmniVoice: Increase "Speed"
Kokoro: Increase "Speed"

The other models don't have a way to directly influence speed. I'm working on adding post-processing directly within VCP so you don't have to swap programs.

Hey Samer! Unfortunately Microsoft doesn't let me re-download older versions :(

It's very unlikely that the accuracy changed between versions so maybe we just have to figure out the settings you used. I'll message you on Discord and we chat there.

Try this:

1. Delete any files you've extracted
2. Right click on the zip file and press "Properties"
3. Check the box next to "Unblock"
4. Re-extract all the files

Let me know if it works!

Try this:

1. Delete any files you've extracted
2. Right click on the zip file and press "Properties"
3. Check the box next to "Unblock"
4. Re-extract all the files

Let me know if it works!

Hey! Can you try running this command from a command prompt:

start %LOCALAPPDATA%\SongCreatorPro 

And delete the "packages" folder and restart Song Creator. If that doesn't work, send me your config.json file and any logs in the logs folder.

Really appreciate the kind words!

I'm actively working on fixing SCP bugs so expect it to work...eventually. Also feel free to send me your logs so I can take a look.

Great success! lol

If you open the data folder through VCP, does it open the C: folder or the other drive?

Thanks! Are you using the latest version (1.3.12)? 

Just added this short video to the main page!

Hey! Good questions and apologies for the broken generations, I've been working on fixes all weekend. I'll start with the differences:

• The repo you linked requires a minimum of 10GB of VRAM. Song Creator works on 8GB (possibly 4GB too) and without a GPU at all. 
• The repo uses a model that does not allow commercial use of generated songs (this is a big one).

And some planned features:

• A full REST API for programmatic generation
• A Web UI for accessing SCP via a browser on other devices
• The ability to fix/improve sections of generated songs Released in v.1.2.3
• Extracting/Adding instruments Released in v.1.2.3
• (Long-term) More/Better models as they're released
• (Long-term) A native MacOS app

My goal with SCP is to make song creation as easy as possible from installation to generation. As you noticed, I only just got v1 stable and it has what I'd consider the bare minimum features for song generation so there's plenty of work left to do.

Having said all that, the roadmap isn't set in stone so if you have any feedback for new features or improvements, I'm all ears!

Hello again!

I just released Song Creator Pro, a tool that lets you use AI to create songs! Works great for game music tracks as a starting point or game jams if you need something quick. The more descriptive your prompt, the better the output. There's also a built-in language model that will enhance your prompt if you're feeling lazy. 

Get a 50% discount if you've already purchased Voice Creator Pro!

I just released a version of my app, Voice Creator Pro, here on Itch. It's a tool for using AI to generate extremely human-sounding speech, either by cloning an existing voice or prompting for a completely custom one. It works great if you're a content creator or game dev who wants to add quick voiceovers. There's also a full HTTP server so you can hook VCP directly into your existing workflows and automations.

I've got a ton of features on my roadmap but feel free to let me know if there's anything you'd like to see! 

Imagine being so dumb you didn't make the files available 😭

They should be downloadable now, sorry about that!

This was indeed a bug! The correct behavior is to unlock the Shaman once you get 32+ poison at one time. Instead of checking against the player's total poison stack I was incorrectly checking against the last stack added. I've fixed it now so we'll push a Catalog update ASAP.

Ah nuts I forgot to change the version number! I uploaded a new file with the correct version number.

🫡

The latest version of Binary Dungeon is FINALLY here with a bunch of new content and bug fixes. 

New Characters

2 new characters have been added: the coin-obsessed Dwarf and the toxic Shaman. The Dwarf's kit centers around adding utility to gold, like using it to deal damage or to duplicate existing artifacts. The Shaman is immune to poison and focuses on increasing poison stacks to heal or do damage. Both characters are locked by default so you'll have to delve into the dungeon to unlock them.

New Artifacts

10 new unlockable artifacts have been added to each biome. This brings the total artifacts to around 22 per biome! Each run should feel even more unique now. As a reminder, a new artifact is unlocked every 100 exp.

New Dungeon

There's a third dungeon that increases difficulty by adding traps to campsites. You also now have the option to start dungeons 2 and 3 with either bonus gold or bonus health.

Quality of Life

Options have been added to the System Menu to restart the run or go back to the main menu.

♥️

Thanks!!

Thanks for the kind words!