@Jinan

Devlog #4 Nothing too important today
After the gym today, I was really tired, yesterday I wroked on auths, watched some Youtube videos and tested token checking and auth systems on my code, it is very close to the point where I can release it in GitHub to be used locally but needs more work to release publicly as an webapp according to my plans.

Spent a lot of time actually making sure the website works, changed the texts to mine which I actually wrote instead of random placeholders, also making a detailed guide page in the side in text, video and images for people who have other disabilities so it stays super accessible!

P.S: Thanks to everyone who is following my project and liking my posts/devlogs! @Staudt @seti @Arsen @GoldenClarinet @253_Mathilde_IYKYK @stratustraipsing @ajmaleee__ @wayne14, truly appreciate y’all! <3

#Devlog3 Local mode has some issues but it is looking good!
My plans for today/tomorrow:

• Make the website look nicer, add a dark/light theme
• Try to make the code actualy clone your voice using XI and then play it
• For some reason, the code bugged and now rooms don’t close and take up server resources unless I stop the site
• Research a bit on how to get the site public or if databases are needed, likely trying cloudflare pages though.

Devlog#2 Code improvements, security and website redesign

• For some reason, the audio was not playing from the sender to the receiver due to some issues with websockets and content security settings, after debugging for 30 minutes, finally came across fixed settings to fix it.
• Had to recode the html logic to properly take the XI Labs API keys.
• So the API keys need to be safe to release this to the public, but currently I have no idea how to keep the keys secure as this is an open-source project, so I took some help from Claude cowork to help me setup security features, it added git-ignore.
• Added checks to if the sender is running the webapp locally or if it is from production, with logs to explain what to do either way.
• Sender rooms to share their voice were active even after the session ended and the sender could not create any during that time, it was really annoying to fix, but I added a system that cleans the room when it is inactive for over 60s to save server resources and let the sender create another one

Working on:

• Finish the logic for room code generation and sharing directly via browser links.
• Proper authentication
• Maybe a login/sign or just distinguish users from XI API keys?
• Still need to actually QA API key security features though

Note: 4 images is a limit here so couldn’t show all new code but these are the main ones

Devlog#1 Initial prototyping and designs
It was a pretty hectic day, I made a local voice assistive communicator tool for a hackathon that was super unpolished pretty much some time ago under a few hours. So I decided to actually work on it myself from scratch as it would help people who face Tourette’s or VCD(like myself). But this time I plan to release this to the public so people all across the world can use it .

Things I did today:

• Recoded the API logic that allows voice cloning and playback from scratch.
• Made a temporary layout for the homepage now, I reused the color palletes and designs from my personal studio site to save some time.
• Tested a ton with ElevenLabs API, websockets, to make sure it can actually play audio on different tabs.
• Began setting up the rooms-system(I named it myself) so it is actually accessible to anyone with a browser and does not need 3-4 different apps or mess with your mic.
• Began adding info to the page <– will add a video guide and more detailed notes(with pics) on the website to make it easier for people to use, but just for now.
• Did a ton of testing to make sure the website kinda works for now, at least the voice cloning, it will take time to figure out how to keep tokens safe while people use this though.
• Made a ton of stupid beginner mistakes in my code such as keeping my whole logic in a single file so now it is a nightmare to edit unfortunately :(. Was able to hopefully correctly resolve the rest but will check on it tomorrow as I am tired today.

Updated my readme taking inspiration from the resources provided by Stardance and my friends.

Devlog#2 Failed a little but back stronger!
I noticed that my website took over 5s for images to load and it was moved from Hostinger to Cloudflare Pages which is free, had to set it up properly. The build broke locally when I made optimizations but now the site is more optimized than ever before!

Big changes:

• Fixed all image loadings from 5s to under 1s. Hero images load instantly now.
• Reduced animation library just a little bit as some things weren’t even being used.
• Forms were always broken with no fallback, added a fallback now on about page, will make proper forms later.
• Every role on our jobs page now has an apply button that works.
• My ceo personal site page was not connected anywhere, but now it can finally be visited by clicking my name on the teams page!

Small fixes:

• Site mainly contained placeholder info but replaced them with real info, mainly dates.

QoL for myself:

• Templatized everything so I can make pages in the future easily.
• Made a small todo for future plans!

Devlog #1 Personal page setup and a ton of fixes!

Main new thing is my own personal CEO page added!

Spent the last two days rebuilding how the site loads:
• Site is now 3x lighter: 4 MB down to 1.4 MB
• Images went from 3.6 MB to under 1 MB, self-hosted instead of pulled from an external service
• Fixed a few images that had been silently broken for weeks
• The logo is finally working. It was hotlinked from a CDN that died, then got stuck behind a bad cache
• Hero images now load immediately instead of waiting on the JavaScript to finish
• Fixed the site fonts. We were asking for a weight that doesn’t exist, so headings were being faked by the browser
• Sharing cheristral.com now shows a proper preview card instead of a naked link

Should feel noticeably quicker, especially on mobile.