Ask HN: What simple web apps do you wish existed? Seeking ideas for sample apps

An interface to preferentially compare and choose between 2 objects, using a Tinder like interface (swipe left/right mechanics). Every year I do a stack ranking of baseball players for my fantasy league. If I had this interface, I could easily setup an ELO comparison function to create my complete list out of a series of easier comparisons.

I have at times also wished I had this for just about any sort of data tagging. Give it an array of json objects, it displays the data and then lets you swipe super quickly through all the objects in the array, semi-randomly choosing which objects to compare. Would need to write the data out as well - that's the product.

This is probably worth a couple hundred bucks to me if anyone wants to build it and open source under a commercial-reuse allowed license. Every year I spend dozens of hours doing this and don't feel like I have done a great job of making the correct comparisons.

When you mention “objects” do you mean a literal Noun type object? Or purely data objext

That is a pretty cool idea.

This is a great idea! -- I'm going to take a stab at implementing this tonight

I need this too for sorting through all my credit card transactions! Swipe left for business and right for personal

Check out https://github.com/antialiasis/favorite-picker The favorite Pokémon picker is based on this and I have used it (and similar ones, based on Tōhōsort) to do similar sorting.

any luck on implementing this one? would be cool to see if pycob can do it.

How's this?

The home page is a setup where you setup your dataframe. If you already have some csv or something, you can read that rather than getting user input.

The app will randomly select 2 rows of the dataframe and display them as side-by-side VS cards. You then pick the winner from the two. Then the app will repeat random selection/winner as many times as you want.

There's a results page that shows the ELO ranking of all the "players" as well as the history of all the "games" between the players.

I implemented not with a "swipe" interface but just with side-by-side cards and you click on one. It wasn't clear to me how swiping would work when you're choosing a winner between 2 objects rather than making a yes/no decision. If you have an idea for how swipe would be useful, I'm all ears.

Live App: https://compare-and-rank.pycob.app/

Source Code: https://github.com/pycob/sample-apps/blob/main/compare-and-r...

This looks pretty awesome! I am seeing an error when I try to rank things - is something temporarily broken?

I don’t believe you should be getting an error. Could you share what you’re seeing? Feel free to submit a GitHub issue or email me at my username at gmail

System Info

I want to open up a webpage and see, for my local device:

- CPU (core count, manufacturer)

- Memory

- GPU present

- Disk size, free space

- IP4 and IP6 addresses

- Ping to different countries main providers

- My geolocation and altitude

- A pin point of that on an embedded (free, ideally SVG, global) map that doesn't need to specify layers/metadata/cities/countries just show me the projection of lat long onto the globe of the earth and the contours of continents

- My current velocity

- Any sites I'm logged in to or identified by

- My browser version and known exploits against it

- My OS version and known exploits against it

- Connection speed ( link capacity, actual up and down, and rates per second )

- Screen dimensions and capabilities

- Output of any other accessible sensors: gyro, bluetooth, wifi, whatever

I created as much of this as I could lazily so far at https://mycapabilities.me (github pages is borking HTTPS for me tho)

are there any good python packages that get you these (at least some of them) already? that don't have a good front end that pycob can leverage?

Thanks! Could you elaborate on your use case? I'm trying to figure out how much of this can/needs to be client-side vs. server-side

I just thought it was cool. You know, you can open up : about this computer on Mac and windows. Why not on web?

jeeze... wrong house, but you got me right down to the street location.

Glad you could use it! I thought the GitHub pages was having an issue

I want to look at the source. Very helpful.

It’s all in the JS of that site

Here's a thing I always wanted: Show big Twitter discussion with loads of subthreads/replies in Reddit/HN "comment tree" format.

I have twitter thread reader webapp. Currently it requires to run scripts locally to load the thread, so not quite as convenient.

https://github.com/python273/twitter-reader

I looked into doing this a while ago and it’s near impossible due to how Twitter classifies/organizes threads (conversationId) and limitations within the API to derive the full tree.

I am working on something similar to a dataframe explorer called data cleaning framework. Would love to talk to some people about it. The idea is that everytime I start data analysis with pandas, I run through the same adhoc data cleaning commands in many configurations. I'm building a graphical UI that lets you quickly try different commands AND it outputs python code to accomplish the same transforms.

https://github.com/paddymul/dcf-server

Youtube channel or playlist to RSS feed, that I can then add to my podcast player. The link should point to the audio version of the youtube video.

There is already one out there that I use, but its very badly designed.

Would you pay for something like this? Sounds like a fun weekend project but saving the files would get costly after a while. I’m imagining you want to use this for channels that don’t offer traditional podcasts but maybe operate in that long format (multi-hour)?

> use this for channels that don’t offer traditional podcasts

Yes. Plenty of such channels out there.

> saving the files would get costly after a while

You don't have the save the audio files. Its possible to just directly link, within the RSS feed, to the audio files hosted on youtube servers. The server simply has to periodically generate the RSS feed using YouTube api calls, and then host the RSS files. Surely, storing even 10s of thousands of RSS feeds won't cost that much. They aren't also being requested that often, so bandwidth costs are minimal.

> Would you pay for something like this?

Sure.

Does youtube host an audio only version? I wasn't aware that was something that they did. My understanding is that all Youtube content is video only. Even if the screen is blank, you're being served some content with a video codec. I've done a little cursory searching and to my knowledge, Youtube does not offer an MP3 link to any content (from their API or embedded on the page).

They offer a mix of audio-only, video-only, and video+audio. Here is a list for a particular video (great music btw)

    > youtube-dl -F https://www.youtube.com/watch?v=GMiaVo_x1EY
    [youtube] Extracting URL: https://www.youtube.com/watch?v=GMiaVo_x1EY
    [youtube] GMiaVo_x1EY: Downloading webpage
    [youtube] GMiaVo_x1EY: Downloading android player API JSON
    [info] Available formats for GMiaVo_x1EY:
    ID  EXT   RESOLUTION FPS CH │   FILESIZE  TBR PROTO │ VCODEC       VBR ACODEC      ABR ASR MORE INFO
    ──────────────────────────────────────────────────────────────────────────────────────────────────────────────
    sb2 mhtml 48x27        0    │                 mhtml │ images                               storyboard
    sb1 mhtml 80x45        0    │                 mhtml │ images                               storyboard
    sb0 mhtml 160x90       0    │                 mhtml │ images                               storyboard
    599 m4a   audio only      2 │   13.42MiB  31k dash  │ audio only       mp4a.40.5   31k 22k ultralow, m4a_dash
    600 webm  audio only      2 │   16.34MiB  37k dash  │ audio only       opus        37k 48k ultralow, webm_dash
    139 m4a   audio only      2 │   21.27MiB  49k dash  │ audio only       mp4a.40.5   49k 22k low, m4a_dash
    249 webm  audio only      2 │   23.83MiB  55k dash  │ audio only       opus        55k 48k low, webm_dash
    250 webm  audio only      2 │   30.99MiB  71k dash  │ audio only       opus        71k 48k low, webm_dash
    140 m4a   audio only      2 │   56.45MiB 129k dash  │ audio only       mp4a.40.2  129k 44k medium, m4a_dash
    251 webm  audio only      2 │   65.21MiB 150k dash  │ audio only       opus       150k 48k medium, webm_dash
    17  3gp   176x144      6  1 │   21.50MiB  49k https │ mp4v.20.3    49k mp4a.40.2    0k 22k 144p
    597 mp4   256x144     13    │    4.97MiB  11k dash  │ avc1.4d400b  11k video only          144p, mp4_dash
    598 webm  256x144     13    │    6.97MiB  16k dash  │ vp9          16k video only          144p, webm_dash
    160 mp4   256x144     25    │    5.51MiB  13k dash  │ avc1.4d400c  13k video only          144p, mp4_dash
    278 webm  256x144     25    │   15.44MiB  35k dash  │ vp9          35k video only          144p, webm_dash
    133 mp4   426x240     25    │    9.53MiB  22k dash  │ avc1.4d4015  22k video only          240p, mp4_dash
    242 webm  426x240     25    │   28.68MiB  66k dash  │ vp9          66k video only          240p, webm_dash
    134 mp4   640x360     25    │   15.47MiB  35k dash  │ avc1.4d401e  35k video only          360p, mp4_dash
    18  mp4   640x360     25  2 │ ~ 73.46MiB 165k https │ avc1.42001E 165k mp4a.40.2    0k 44k 360p
    243 webm  640x360     25    │   55.53MiB 127k dash  │ vp9         127k video only          360p, webm_dash
    135 mp4   854x480     25    │   22.52MiB  52k dash  │ avc1.4d401e  52k video only          480p, mp4_dash
    244 webm  854x480     25    │   89.33MiB 205k dash  │ vp9         205k video only          480p, webm_dash
    22  mp4   1280x720    25  2 │ ~ 90.62MiB 203k https │ avc1.64001F 203k mp4a.40.2    0k 44k 720p
    136 mp4   1280x720    25    │   32.23MiB  74k dash  │ avc1.4d401f  74k video only          720p, mp4_dash
    247 webm  1280x720    25    │  166.48MiB 382k dash  │ vp9         382k video only          720p, webm_dash
    137 mp4   1920x1080   25    │  106.40MiB 244k dash  │ avc1.640028 244k video only          1080p, mp4_dash
    248 webm  1920x1080   25    │  271.78MiB 623k dash  │ vp9         623k video only          1080p, webm_dash

I was looking for something similar, but figured the use case was so narrow that nobody else was interested- what are the issues with the solution you found? Thinking of building this out.

Its something someone wrote a long time ago, and it seems to be on its death bed.

It doesn't create feeds out of playlists iirc. Sometimes on a large channel, you just want to listen to one particular series.

It doesn't let me choose the feed icon/image. Instead, the image is the text of the service name. This is very annoying because scrolling through my subscriptions on my podcast player, I primarily use the image to identify the one I am looking for.

how would you implement this? just take in a URL that has an RSS feed and then link each item on the feed to the podcast player?

Each item in the RSS feed links to the audio of a single episode (with the item title/description set as well) of the podcast. Podcast players can subscribe to the RSS feed, and display episodes as usual.

I have two that I've been thinking about that I've been too lazy to cobble together myself:

A page change tracker that A: pumps the page through uBlock origin before each snapshot and B; produces an RSS feed with a summary of changes.

A web page to epub converter that A: sends the page through uBlock origin or similar, B: sends the result through the postlight parser, and then C: gives the user a chance to customize the result before packaging it up as an epub. Customisation would include setting a cover image (or none,) setting a title and other metadata, choosing to include a TOC or not, renaming any chapters if a TOC is included, etc. Currently my workflow for this involves someone's "web page to epub" converter personal project that isn't entirely designed for single pages and a bunch of time on Sigil cleaning up the resulting output.

what do you mean by uBlock origin? what's the use case here?

I've always wanted to build an "I told you so" webapp.

Something that would store phrases associated with people with an incremental counter. Obviously the list of phrases, people, and counts are unique to the user.

This way I can say to my daughter, "I've told you <checks webapp> 46 times now to clean up your room!"

Dumb, but fun :)

Try this approach:

Don't clean your house ever, start leaving your dirty clothes outside her room, don't wash the dishes, don't take out the trash, let it smell.

Make your daughter be the to tell you "I told you 46 times to clean up the house!".

This is getting dark

Why not just burn the house down while we're at it

Maybe also -

Add an immutable time stamp to the phrase, so I can say “Alice warned Bob about that problem on DD-MM-YYYY” so that when Bob gets hit with it 6 months later, exactly as Alice predicted, Bob can’t say “You never told me that”

As someone with a near perfect verbal memory (I can recall conversations verbatim from my whole life) it’s incredibly frustrating to have people deny that I told them something, when I can recall the exact date, time and words, when I did :P

Voice activated calculator that allowed me to rapid fire simple math from across the room.

If it did math while people were meeting and conversing and spat out an answer, that would be pretty cool.

I also like counting apps: take a picture of a bowl of Cheerios and it counts the 'O's, or the number of pipes in a rack at a hardware store.

I'm sure they've been engineered already or at least thought of before, but so far voice calculators seem slow and clunky like voice assistants, and counting apps lack surety and I wouldn't bet anyone's job on them that I've found so far.

can't google assistant or alexa do this? why would you need a web app?

https://pypistats.org/ tells you about pypi fownload stats

oh this is cool

Markdown articles to JSON and/or JSON upload to websites.

That's a soft JSON not a must-be JSON. I'd kill for an all markdown workflow where I could build content/blog posts in my markdown app (Logseq) by selecting a layout template, writing the articles and selecting pictures, and then I could kick off a simple upload process to my web server.

That's a great idea! Is there a specific JSON format you'd need?

The candle shines the light on my ignorance, JSON is JSON to me. Use whatever you deem best if you decide to pursue this!

How does this look? I'm happy to make changes to suit your needs:

Live App: https://markdown-to-json.pycob.app/

Source Code: https://github.com/pycob/prototypes/blob/main/markdown-to-js...

Slick nice!

holy! 42 lines of code - love it!

This one smells a bit like something I run into at work sometimes, where a non-technical person makes a technical decision and the technical people don't sufficiently challenge it.

If you're trying to convert markdown documents into webpages, the most likely output format would surely be HTML, or perhaps something custom to the site like MediaWiki markup.

It's totally possible that a site would allow for new documents to be uploaded in a JSON format, but the format would have to be specified (e.g. which keys are used for the post body and subject) - so "whatever you deem best" is unlikely to work, it would need to be "whatever my webhost expects, which is documented -here-"

I'm happy to be wrong here, and zainhoda's markdown to JSONified HTML is interesting regardless - but I suspect you really wanted a markdown to HTML converter. ex: https://markdowntohtml.com/ or something more extreme like a static site generator: https://www.getzola.org/

Not necessarily. With the push towards headless development, using JSON to fill content opens up layout options significantly. Being tied to pre-made HTML runs you into significant issues trying to do anything that isn't just markdown page presented with HTML.

I left the decision to him because why interfere with someone else who probably has a stronger handle on their own strengths. As for your webhost concern, I can always change webhosts.

Great Idea. I have a usecase where I have more than 15 API's and several Postgres tables. I want to quickly create a Web UI with form for support teams to perform GET/POST/DELETE/PATCH requests. And also show the table contents.

Web UI with a form + Table Exploration from Database would be great.

the pypi analytics one is cool. what about doing something similar for github analytics? i find it more interesting and expansive than pypi.

Something with LLMs that uses sanitized data so it’s useful and not jailbroken and shut down within a month.

For example:

- an app that forecasts the weather where you are in 1 sentence.

- an app that summarises user reviews of points of interest around you.

- an app to summarise YouTube videos based on their transcripts.

- an app that describes your address in 1 sentence as you would to an emergency service (“26 miles West of Placename on E25” or “on the corner of Eve drive and Redmond Avenue”, or “at Trafalgar square”). Shown with a map with all used place names for confirmation.

- an alternative history daily news generator with several settings (cyberpunk, soviets won space race, world peace, world war 3, Nazis won WW2, the Earth is flat, brave new world, etc).

- an endless text based adventure game where AI gives you some story and a verb each turn. You can set a theme in the beginning in 40 characters.

- an app to transcribe and summarize voice recordings, for lectures and journaling (should be sanitized for jailbreak instructions but would be a useful app).

- an API or app to generate long memorable pass phrases (based on semi random context like a random Wikipedia page).

> - an API or app to generate long memorable pass phrases (based on semi random context like a random Wikipedia page). Somewhat close - see https://www.correcthorsebatterystaple.net/ for relevant xkcd https://xkcd.com/936/

I was thinking along these lines, but we could remember much longer passphrases if they made semantic sense - the words were not random.

For example, here's what ChatGPT comes up with when asked to create some short rhyming passphrases with symbols or numbers:

* "Not all who seek, find."

* "Speak with kindness - inspire greatness."

* "Wise to listen 1st before speaking."

How about this entropy? These are the final bosses of passwords.

The book summary! But curious around costs for that

Cool idea, I have some medical ai applications

interesting, like what?