I have a lot of ideas. Most of them are reasonable and, I think, meager — my mind gravitates towards things that are:
This is my actual list of project ideas, not some sanitized list for public consumption. If I start something new, it's going to come from here! Please take these things; my desire to build them is far outweighed by my desire for them to exist. (And if you do, email me so I can play around with it.)
(Along that whole actual list, not sanitized list thing — some of these might be a bit inscrutable. Apologies, but just ask for clarification!)
A hosted and rated instance of queryparser, which is a very useful tool for data organizations that very few people use because standing up and deploying Haskell is incredibly annoying.
Hard copies of PEP
ListenNotes for newsletters. In general, I think the metadata & infrastructure layer for newsletters is very much in its infancy: similar to where podcasts were a decade or so ago.
Plaid for DNS providers. It's incredibly annoying to manage DNS records for SaaS apps like Buttondown where you need to ask users to set records; even very technical ones will run into esoteric issues. 1
A simple app that hooks into your GitHub repository, parses your Pipfile/Gemfile/etc, and emails you periodically with information about new major and minor releases that are of interest to you.
A microsite that is an easier and more interesting display of the biggest tech mergers of all time.
A blog for "pro tips". My former employer had a Slack channel for this — aptly named #protips — and the number of very small but useful things I learned from lurking there was incredibly high.
A blog (or network of blogs) centered around easy writing & reading of friction logs.
“How much does this meeting cost” hooked into levels.fyi
An analytics & insights layer on top of GitHub Issues.
A nicer way to find your representatives and their voting history.
An app that, given an image, pulls out the fonts used in the image. 2
An app that is like blank.page but auto-appends the timestamp for taking meeting notes. 3
A local-first web app for tracking periodicity-heavy workouts like SBS 2.0.
A basic SaaS app wrapper around Google's Postmaster APIs. 4
A safari extension to force you to add accessibility before tweeting images
Locket.camera, but for sharing your calendar with someone. Me and my fiance (and I'm sure countless couples around the world) have a morning routine of texting each other a screenshot of our calendar; this would ameliorate that.
A more pleasant version of TwitchTracker, the data for which is freely and ergonomically offered by Twitch.
A more pleasant version of all of these online script sites, of which subslikescript.com is the most notable but certainly not the only.
A site that sits as a tracking substrate atop all of the public job boards out there. I'm not sure how this could be immediately useful, though I suspect that it would be (and probably very easily monetizable; I have spent enough time orbiting the engineering recruiting space to know that if you can make a single placement 1% more likely to happen, you are worth a lot of money.) This is also an attractive project by virtue of how fun it is to scrape things and apply a lens of structured data atop it. I don't know if there's a particularly useful consumer angle, but I have a feeling you could at least do a lot of good content marketing.
A micro-app that allows someone to share a CSV, which is prettified and persistently stored. There might not be sufficient ergonomics here to warrant using over something like Google Sheets, but there are a huge swath of "CSV viewers" littering the internet and the vast majority of them are poor, either due to bad experience or a proliferation of advertisements. (Shout out to Tinysheet, which is quite nice but lacks some of the features I want.)
A better RQ dashboard. I use RQ for a number of various purposes: it is a Python library for asynchronous job execution very much in the vein of Sidekiq or Resque. It ships with a dashboard, but this dashboard is (and apologies to the maintainers of the package, to whom I am deeply grateful) not very good, at least in my experience. It is adequate and showing a raw count of jobs in the queue, and fails (often loudly!) at every other task. I imagine that you could fairly easy productize this in an open-core model.
A JRPG using Match 3 mechanics. I'm playing Hero Emblems 2 right now and it's a perfectly cromulent game, so to speak — it is a nice smash-up of the two aforementioned mechanics. But it's very bare-bones on both sides of the spectrum and a poorly engineered/balanced game to boot — but despite that it's still quite fun, because Match 3 solves one of the core problems in JRPGs (boring, rote combat). Someone could make a really fun, good version of this game with a Threes-esque emphasis on polish and gameplay. And they wouldn't need to be a hardcore game developer! (I promise I won't start building games.)
A productivity app aping the concept of "daily quests" as popularized by MMOs. Basically, I want to be able to input a list of a dozen or so non-trivial tasks that I tend to not prioritize and are too large/cumbersome to do every day ("watch a movie", "publish a new blog post", et cetera) and have an app select
n of them that I do every day. There would be a streak component (so the goal isn't to do all tasks every day, but to do the full randomly selected subset of tasks every day). This is probably a very small niche, but it's also such a small surface area for an app that you could probably charge $1.99 for it and make a few hundred bucks a month.
A dedicated comparison site for infrastructure providers. There's a bit of a public and long-term exodus from Heroku over the past six months that I expect to accelerate and continue over the next two years, and many of the decision makers leading that exodus don't quite know where the landscape stands. Referral bonuses on such conversions are very lucrative, and also i think this is a fairly underexplored content space.
A simple app that lets you track various conversion rates for Stripe Checkout sessions and Payment Links, broken up by various UTM parameters. As of the time of this writing (Fall 2022), Stripe offers conversion funnel analysis for these things based on some very specific pre-baked pivot points (region, device type) but there's little data portability or extendibility. I think this is a very small app to build (on the order of weeks) and while the market size is not huge, there are many Payment Links users who are technical enough to know what UTM codes are while not technical enough to be able to roll their own solution. (You'd have to worry about Stripe building this functionality into their core app, but again the startup costs for this project are so small I think the value proposition is still worthwhile.)
An open source (or closed source!) implementation of one of my favorite internal Stripe tools.
Literally just this, as a product or blog
A micro-app for creating DM-like videos like this one
A Chrome plugin that adds OpenAI autocompletion to any textarea on the page.
Conversely to the above point — an API for detecting whether or not a given input string has been generated by an LLM.
A MicroSaaS that takes a repository of issues (Linear, GitHub Issues, etc.) and turns them into an easily printable PDF of index cards for triaging.
This exists now, and I am thrilled to be a flagship customer. ↩
Googling reveals that this already exists, but simply does not work. The web is littered with tools that were made ten years ago that can be trivially made much better and faster. ↩
Twitter user @joodaloop took this idea and ran with it to great success. ↩
In general, I am bullish on the genre of app that is "take an annoying but incredibly useful API offered by Google and make it pleasant to interact with." For example, Briefmetrics. ↩