milo1004

In this 6 hours of work, I’ve added these:

• AI response streaming (finally!!)
• timer precision fix

I will continue to fix the stopwatch precision issue on the next devlog

In this hour, the most I did was fixing, cleaning up the backend code, and increasing the border radius and reducing the border size to make the website less bulky

• Fixed weather fetching error from backend
• Fixed image not loading (because i forgot to add .png to the name) in weather fallback mode
• Cleaned up backend AI code

In this hour, the most I did was fixing, cleaning up the backend code, and stylizing the widgets

• Fixed weather fetching error from backend
• Fixed image not loading (because i forgot to add .png to the name) in weather fallback mode
• Cleaned up backend AI code
• Increased border radius of the widgets
• Reduced widget border size to make the website less bulky

In this ~8 hours of work was quite a hard time. But I’ve fixed almost everything in the timer and stopwatch boxes.

• Removed fade in animation when opening utilBox
• Fixed ridiculous numbers in stopwatch after switching from timer
• Added Markdown support to the AIbox
• Added simultaneous timer and stopwatch sessions
• Fixed unexpected disappearance when utilBox buttons are pressed

In this 1.5 hours of work, I’ve added the logics for the timer. It was quite hard, but it worked out anyways.
My next goal:
Add stopwatch logics
Add focus mode UI and logic

It’s gonna be harder, since I need to manage the storage caching.

The timer can also be paused, or reset.

In this 4 hours of work, I’ve worked on:

• Added the UI of timer utilities
• Added logics to the timer
• Added a stopwatch UI (Logics are not yet implemented)
• Fixed some more overlapping issues in widgets

Upcoming plans:

• Add logics to stopwatch
• Make timer/stopwatch sessions persistent after reloads
• Add logics to timer buttons

In this ~14 hours of work, I’ve worked on:

• Switching Piston AI back to a webapp instead of a Desktop app, to make it cross-platform and reduce the trouble of building for every single architecture for Win, Mac and Linux.
• Worked on timer utilities that lets you set timers and stopwatches in case you need them
• Planning on adding focus mode, which once activated, the only things shown on the screen would be:
  A black screen, a big clock shown on the screen, and ur set timers or stopwatches. Might add some other info to the focus mode, but the goal is to keep the user focused on one task.

Bug fixes:
Automatic width/height change when some of the elements overlap with the edges, to ensure that the elements don’t get cropped by the edges.
Other than that, the “create task” button has been changed to a static position (next to the title of “To-Do”) to avoid awkward positions like outside of the todo box.

I’ve just started this new project to learn a SolidJS.
In this commit, I’ve added a simple OOBE stage, which is still in progress. Will release an OOBE setup page in the next update

I also spent many time figuring out how to place the UI elements, but it’s. overall enjoyable!

This time, I’ve updated the AI box to use speech bubbles instead of pure text, since there have been complaints about the ugliness of the old versions

Also, I’ve built the Linux version of Wails so ALMOST everyone (well Linux ARM64 haven’t been built yet) can use Piston AI locally!

The reason that this devlog is 36 hours might seem impossible, but it’s an issue from FlavorTown itself, that it deducted ~36 hours from this project, and this devlog fixes it by adding back the 36 hours. This is not a fraud, and was verified by a FlavorTown staff.

After many many hours of work, I’ve finally successfully turned Piston AI from a website to a real webapp!

The reason why it took so long, is because of Tauri. I tried many ways to just embed the image files, changing the src.js to src.ts, and many other things.

But, I’ve decided to scrap the Tauri plan, and use Wails instead (uses Go instead of Rust). And in just 2-3 hours, I’ve finally embedded all the image files and modules to Piston AI with a working fetching system between the backend and the frontend!

What I will be working on in Piston AI, is creating a introduction page for Piston AI, for downloading the Piston AI binaries and also moving the original Piston AI website to a sub-page which will no longer be maintained since I’m moving on to the local version.

And compiling for Linux

Now it’s prepared to be shipped as a binary!!

In this ~4.5 hours of developing, I’ve added

• a new font to Piston AI, which is Jetbrains Mono!
• data management, e.g.: Export, Import data and Erase data. Can be found in the settings tab
• Fixed weather fetching mechanism on backend by adding CORS headers and using Workers as the proxy server
• Added credits to sources like wallpapers and fonts to README.md

Below are the official GitHub commits:

• Update weather fetching to use POST method and display current weather data
• Add data import/delete UI and update worker endpoints
• Added data management
• beta update: added data management
• Completed the todo list feature

In this new update, I’ve completed the todoList, which I was working on before. You can now add any tasks pending! You can also create a task with a deadline, so you’d actually know you’ll have to finish that task by a date! It also will delete itself after the deadline has passed.

The most important part, is that all the tasks can be viewed by the AI, so you can literally ask it for guidance if… the task is homework lol

I’ve finally added the backend files to GitHub! It actually took a lot of time (about like 1.5 hours) since configuring Wrangler is PAINFUL!

And for the rest of the 1.5 hours, I’ve added a TODO list to the website. Also, no logic yet, and UI only (also painful for me to configure the position of the box). I also removed the alarm feature, since web-based alarms are… not convenient. You won’t see the box yet, since I’ve hidden it, but you can figure out the code and make it visible yourself!

The image below is the version of a visible TODO Box, which is just a preview (do this by pressing F12 to the dev console, make it in a separate window, type in
todoBox.style.visibility = “visible”
and hit enter. After that you will be able to see the box! (Bare minimal UI lol)

In this 3 hours of work, I’ve enhanced animations. But, most importantly. I’ve added the memory management back to Piston AI!! Now, just ask Piston like “Remember that my locker key is in my wallet”, and it will add the memory to your localStorage!!

I’ve worked on a theming function in Piston AI! Now, you can set wallpapers freely, instead of a dull black backdrop.

In this new update, I’ve made the hardcoded “Welcome back, (username)” to better greeting words like “Hey”, “Evening” and others, depending on the time!

I’ve also moved the JS code from index.html to an independent file “src.js”!

Btw the 23hrs of work, some of them is because I forgot to add them to this project when coding in Vim. It’s not a scam.

In this new update, I’ve added a quoting system (quote of the day), so every time the page refreshes, you get a valuable quote!

In this new update, I’ve been working on a new feature, the alarms function. Just like the older Python builds!! But, it’s still in-dev, so there’s only UI elements, not the logic and stuff. (Haven’t linked them yet)

Ofc, since this is a BETA build, I haven’t deployed the files to CloudFlare as a website yet, so stay tuned! Check the beta version out by downloading the source code from GitHub, and I’d recommend you to run it locally with Live Server (VSCode)

Super Huge Update!!!!!
I’ve remade, basically every single thing. Everything is now shifted from Python to HTML, JS and CSS, with CloudFlare’s Workers AI as both STT and the main AI assistant! This is not the most stable version yet. I haven’t implemented all the commands, but telling the weather, time, and date, as well as clearing the chat works for now.
Btw, a container would pop up after you ask the AI. Click the container to minimize it.

In this update, I’ve added support for macOS, and also persistent memory. It is different from history, which you can tell the AI to save data persistently.

I’ve replaced the plain STT engine with MicroWakeWord as the WakeWord engine. It’s more accurate, and prevents mis-detecting.
To train the model, you will first have to do “python setup.py”, and then it will guide you through training the model. You can either record “Hey Piston” as the wakeword, or choose your own wakeword. Just record your desired wakeword as it prompts you to.
Btw, don’t start saying the wakeword until “🔴 Recording… press ENTER to stop” pops up
And also, I’ve coded the script to abort the recognition when no one is talking. But if you said something, it would minus one in validCount, so the chat can go on.

In the previous 4.5 hours, I’ve rewritten the these things:

• main.py -> To be less buggy
• executeMessage in utils.py -> To make the execute smoother
• AlarmGoOff.py -> Just to rewrite, and now it works
  …

And yeah, I’ve also fixed the alarm.set.py, and tested to be successful. For AlarmGoOff.py, I’ve made it able to say the purpose (“label”) of alarm, and play a ringtone. Ringtone credit: pixabay.com

here’s an example format of an alarm in data/alarm/alarms.json. Feel free to edit anytime!

In the previous 7.5 hours of coding, I’ve added an alarm daemon, a weather app, and Piston’s main speech-to-text engine “VOSK” has been replaced by OpenAI’s FasterWhisper.
You can simply tell the AI to list, add and clear alarms in a natural way, e.g. “Wake me up at 8 in the morning.”

You can also ask the AI to tell you the weather like windspeed, temperature, and wind direction in your region.

For the STT model, it is now more accurate, and the recognition doesn’t suddenly stop when you’re still talking, by using the RMS technology to detect silence from the mic.