Hacker News new | past | comments | ask | show | jobs | submit login
Show HN: Build live location sharing in your app with HyperTrack (github.com/hypertrack)
103 points by arjun27 on May 20, 2017 | hide | past | favorite | 40 comments



I highly recommend also Open Sourcing the server, or else you'll have to compare against Open Source alternatives:

- We did a prototype last July on saving 100M+ updates for $10/day all costs (cpu, disk, backup). Proof: https://www.youtube.com/watch?v=x_WqBuEA7s8

- You can 1 click deploy to a free Heroku server here: https://github.com/amark/gun#heroku

- Watch a demo of the GPS tracking here: 1min https://youtu.be/7ALHtbC9aOM

- Demo of the app here: (warning: asks for location) http://gps.gunDB.io/

- Source code here: (only ~200LOC, lots to improve) https://github.com/gundb/gps


Pricing looks way too expensive to me even for small to medium scale usage. And if my business cross the medium scale , then why wouldn't i really build it inhouse. Seem like a classic case of product looking for problem.


Its 4 cents per action. An action here is one trip which would involve multiple updates. Doesn't seem that expensive.


Right. I use doordash all the time, and pay $5 - $10 for delivery. It gives me occasional status updates on my delivery, but no real time tracking. I'd love to see this in doordash, and if it costs them $0.04 from the $5 fee it seems like a no brainer. Same for so many other similar services.


IIRC, doordash used to provide a live map a couple years back. I'm guessing they have all the data and could share it again if they desired. Presumably they have some product reason for not offering it.


They still do around here (Boston, MA).


The question is why should I even pay you $1 cent for a service for which existing underlying APIs are already extremely good( android or iOS location API and Google Maps API )


We have several large scale businesses that are paid customers and find it cheaper to use us, than build it in-house. It's about the cost of an engineer, and that's a fraction of what they will incur, ongoing operating costs notwithstanding


Businesses for whom location tracking is critical process component and not a good to have such as logistic companies or lets say food/medicine delivery companies who also already have an app deployed with their end agents, the additional cost for them to get location tracking and reporting integrated in their mobile apps is trivial and also there is no technical moat as is there in lets say messaging or SMS APIs ( like Twilio). Either I don't understand your customer set, or problem you are solving is simply not big enough. You are trying to do to location tracking what twilio did to messaging, but reality is location tracking as a service doesn't make a very solid business case here. But best of luck :)


"... the additional cost for them to get location tracking and reporting integrated in their mobile apps is trivial..."

As a professional developer I would say that its not trivial, at least not cheap to develop. Theoretically all problems you know how to solve are trivial. Professionally you know that developing a solution that is flexible, reliable, supportable, extendible, user friendly and maintainable takes time, expertise and money to develop.


We've been amazed at how developers across logistics and delivery companies, messaging products, SaaS platforms, and marketplaces have used our SDKs and APIs to build live location features in their apps, and we're very excited to enable all of them. Thanks for your wishes!


It'd be easy to say that about an SMS API too though, no?


To what extent do you promise your customers that their clients' data will be both well secured and remain private?

Providing a middleman with this kind of information is a little too much trust for me to think in terms of cents per action.

Spying-as-a-Service, haha!


Still too expensive. Instead of biasing your user studies with big companies that already use your product, you should listen to the greater population of potential users. But hey, that's startup 101 and I don't expect you to know that.


Interesting I developed very similar application please take a look at this app https://play.google.com/store/apps/details?id=com.spacetime


Didn't Google Maps do this already for free ? Why should I abandon those and use this ?


Google Maps and other map APIs are great for visuals, routes and ETAs. Where they lack is generating a battery efficient accurate location stream and mapping that to your app's workflow. Plus, map APIs are priced on the basis of number of API calls or views, both of which can lead to unpredictable cost for such a feature. For a more detailed analysis, check out our blog post: https://blog.hypertrack.com/2017/04/23/building-live-locatio...


This is pretty cool - https://blog.hypertrack.com/2017/04/18/pitfalls-using-locati...

Is there something in the protocol that makes it more efficient?


I'm guessing you're referring to MQTT. With smaller packet sizes it's more battery efficient than HTTP if you're making continuous network requests. Great comparison post here: http://stephendnicholas.com/posts/power-profiling-mqtt-vs-ht...


How much of a battery drain is it to be constantly GPS tracking?


<1% per hour. This is achieved by: 1. use consumption patterns to know when to transmit data (eg, if no one is actively tracking, no need to transmit data on radio) [1] 2. using accelerometer and gyroscope data to know when to collect data (eg, if you're not moving, no need to use GPS)

[1] We blogged about how we built out our variable frequency model: https://blog.hypertrack.com/2016/11/28/battery-efficient-rea...


The blog post claims ~5% per hour.


Ah, old post actually, dated Nov 2016. Just wanted to share it for the variable freq approach - sorry for the confusion! We released a major update of the SDK this March which brought in significant improvements due to sensor data.


Do you work for Hypertrack?


Yes.


Looks pretty cool! Do you have something for iOS as well?


It's in the works Parul! Should be out within a month


Is Hypertrack open source?


No, this is just a demonstration of how to use their API in an app.

According to the pricing[1] page, actual use is US$0.04 per "action" (which they describe as "deliveries, pickups, dropoffs, visits, appointments or anything you may decide")

[1] https://www.hypertrack.com/pricing


Maybe I don't understand it but that seems really expensive.


Show HN: How to use Michael Jackson samples in your music

Step 1. Download and install Audacity and LMMS

Step 2. Get clearance agreements from the Jackson estate

Step 3. ...


There's OwnTrack, which is open source and has Android and IOS clients.

https://github.com/owntracks


The mobile SDKs that you integrate in your apps are open source.


What value does this add which Glympse or PathShare already don't provide?


APIs! HyperTrack is built to integrate into your existing product experience and workflows.


What do you mean by API?

Both Glympse and PathShare have fairly elaborate set of APIs on iOS and Android. Not sure I understand the distinction.


My bad, I should have been clearer. Both Glympse and Pathshare are great products and have APIs, but their approach is not developer first. For example, Glympse has guidelines for developers to showcase Glympse branding[1], something I personally as a developer don't appreciate, especially when some of the views are consumer facing. Or the fact that Pathshare's SDKs[2] haven't been updated in more than a year, even though their app has had multiple updates recently. As a developer tool, we focus on enabling developers to build location features in their apps, with the latest and greatest of our product running silently in the background.

[1] https://developer.glympse.com/docs/core/guidelines [2] https://github.com/freshbits/pathshare-sdk-ios


Thanks. Good luck!


could this be used to roll your own uber?


Yep. Location based assignments and live order tracking are two of our most popular use-cases.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: