philwall192

Sesame Shortcuts (Nova Launcher partnership) | Android Development | Part or Full Time | REMOTE or Bay Area

Hi,

If you're looking to not have a boss and work on a cool project that makes money, this might be it.

In July 2017 we launched Sesame Shortcuts in partnership with Nova Launcher. Since then we've become a top grossing Play Store app, with 500k+ installs in 5 months and >20% long term daily actives.

Announcement: https://help.teslacoilapps.com/sesame Press: http://bit.ly/SesamePress App: http://bit.ly/getSesame

You’d be joining an experienced 2 person team. We have a cool roadmap for 2018 and the opportunity to grow outside of Nova. We’re looking to hire a long term fit who can accelerate development. Salary, revenue & equity compensation available depending on experience level and fit.

Relevant skills: - Deep Android knowledge - Nice to have: server, API, and performance experience

Mostly we're looking for a good cultural fit who communicates openly and delivers on time. If you are less experienced but highly motivated... we've all been there, feel free to reach out :)

Email me (Phil) at phil@sesame.ninja. Please include relevant experience, links to work samples, availability, and why you want to work with us. Cheers!

Some crickets in here...

Nobody thinks reading data off the Accessibility Service, building a library of intents and auto constructing user specific shortcuts is cool?

The core idea is to break functionality out of native apps, then build a single interface that can get to things faster.

It uses the Accessibility Service to observe actions, and then Intents and APIs to replicate them. That way we can automatically make shortcuts to stuff you actually use. This clip is a pretty good example: https://www.youtube.com/watch?v=DqGS1kswIlM

Right now we have it working for:

* All your apps (ranked by how often you use them)

* All contacts w/ 1 touch to text, call, or email (WhatsApp too)

* Websites you visit (but not in incognito)

* Music library and new stuff you listen to (Spotify only atm but expanding soon)

* You can Quick Search: Maps, Waze, Netflix, Spotify, Google Music, Chrome, Play Store, Yelp and YouTube

* It makes shortcuts back to your searches

* Hangouts conversations, including groups

* Alarms you set

* Reddit (shortcuts to your subreddits)

* Uber/Lyft see live ride options w/ ETA & Surge Pricing

The interface is a command / GUI hybrid. It’s best feature is the keyboard. We modeled our search off the command palette in Sublime Text. It understands shorthand. If I want to play music by Bob Marley using Spotify. I can type “S ,B, M” and out of the thousands of choices "Spotify: Bob Marley" will be on top. It also learns from what you choose. So “S, B” might work the next time, then just “S”. Like this: https://www.youtube.com/watch?v=ixZ6iB1aRS0

The reason we’re building around the keyboard is that it’s an incredibly versatile tool that is already memorized by users. It really only takes 1-3 taps to get to anything. It’s an adjustment to change your behavior from “swipe pretty picture to enter your phone” to “make a decision now using this power tool!”. But our analytics shows that ppl who take the time to learn the keyboard seem to love Sesame and keep it.

There’s also a ready-to-tap list of your apps. This is more familiar to most users. We wanted the ranking to be based both on how often you use an app how recently you’ve used it. When you open an app it gets a “vote”. We split the day into 15 minute buckets and look at 4 weeks worth of data. We degrade how much each vote is worth based on how far its bucket is from the current time. We give the current bucket a 10X multiplier so anything you just opened is likely to be at the top of your list, for at least 15 minutes. With all this learning, it takes a little time to get good. We’ve gotten a few requests to let ppl pin their favorite apps, but when I follow up a few days later they normally say “nah, it’s good now”.

Building a stable lock screen has been the hardest part. We kept the system security in place. It seemed like the only way to really protect the phone. But it brings a huge cost. It’s not supported by Android, the OEMs all have access to tweak the code too, and different user settings multiply the possible conflicts. The result is we can’t consistently control across all devices how Sesame transitions to the system security screen. We technically sit on top of the real lock screen. I’ve put in every hack I can find or figure out and am happy to share them. I would love to hear from ppl who have dealt with this beast too.

You can open Sesame (see what we did there?) by long pressing the Home button too. This was easy to implement and gives you a way to use Sesame even if you have Lock Screen conflicts. http://stackoverflow.com/questions/14233330/replace-google-n...

We use the Accessibility Service to make new shortcuts for you. Back to the Bob Marley example. If you don’t have a shortcut for Bob yet, but go into Spotify and do a search for Mr. Marley. Sesame will catch that event and be able to make a new shortcut.

The Accessibility Service is sorta this untapped miracle of user data. Right now we use it to do a few specific things. But we tried to build a program that would parse the stream of Accessibility Data so we could map pathways through native apps and automatically make new shortcuts w/o writing new code. It was a combination of graph theory, trees and sets, and Dijkstras pathfinding. It was kind of crazy but I swear we almost got it working before we decided it was going to be the death of us :)

Anyways, we’ve put a lot of work into Sesame and are honestly more comfortable building than promoting. So we hope y’all try it :)

I’m happy to answer any questions and would love to hear feedback.