Apply HN: Moonlighter – Automatically track projects from computer usage
Moonlighter frees you from using a manual time tracker. You no longer have to remember to start/stop the timer, switch projects, or annotate the current task. It pulls information from your computer (file edits, open applications, browser URLs, etc) and connected services (github, gmail, google docs, etc) and uses machine learning to learn which activities are a part of which projects. All tracking is opt-in only, optionally fully encrypted (although this degrades some features of the user experience).
Right now the focus is time tracking, with plans to expand into project management/reporting tools.
Why (did I build this): I had a day job (well I was a grad student), moonlighting on the side to pay the bills. I always had trouble remembering to track my time accurately, and even more trouble estimating hours for a project. Although I don't bill hourly anymore, I know that accurately tracking my time is essential to preparing a good software estimate.
How (are you going to make money): Initial market and customer interviews are freelance developers/designers, although I've had really good discussions from a number of lawyers.
It uses a B2B SaaS subscription model. The Price will start off lower than advertised and increase as more features are added to the web application.
Where (are the competitors): RescueTime is probably the most closest competitor. They have a lot of similar data, but it is less granular and not currently used for project tracking.
DeskTime is similar but also granular, and especially invasive for individual contributors.
There are plenty of other time trackers/invoicing tools in this space, but many of them simply give the problem a pretty face: Harvest, Toggl, and myriad others. Getting detailed data for automatic project/task annotation and project estimation will be a big undertaking for these.
[2k char limit, more below]
14 comments
[ 4.2 ms ] story [ 42.4 ms ] thread- Jeremy Jay (that's me!), I have a PhD in Comp Sci/Bioinformatics, but also do freelance software development. My research covers data integration and information/graph theory analysis. My freelance work has ranged from large software system integrations, a few different startup MVPs (both successful and not). So I have experience with the problems of freelancers/small teams, and experience in data analysis to solve these issues when I have the data available.
- [Need a biz dev/sales person here] - Freelance has given me some experience in this arena, but having a cofounder here would be much better.
When (did I start):
- I currently have a working prototype for OSX and trimmed down prototypes for Linux and Windows. These are installed desktop apps which send data to the main web application.
- I've been working on this part-time for essentially a year. Did a soft beta launch last year, but it surfaced some architectural issues when users and data variety were outside of my expectations. So I'm using that data to improve things and chugging away to get it shaped up for a beta relaunch soon.
For now, I think it's more important to focus on making a great product and start bringing in some revenue. Like I said, I can make ends meet in this area, but it'd be nice to have someone dedicated to that side of the house.
Do you have any questions or concerns? Any feedback is definitely helpful for shaping the product.
2) Let's say I open up Gmail, spend about an hour writing 5 emails to 4 different people about 3 different projects. Could you walk me through how Moonlighter identifies how much time I just spent on each project?
2) The Gmail integration is still in development, but the basic idea is that if you've emailed someone about a project in the past, it'll use that relationship to guess which project you're working on at the moment. The more supporting information it has (e.g. open documents or other stuff you've looked at recently) the better these guesses are.
It does all this on a minute-by-minute basis, so even if you're sitting and thinking about something for a few minutes it'll pick up the little spurts of activity and put the whole block of time into one chunk of activity.
We'll start with a CSV export that will work with these platforms, then further down the line do full API integrations.