Hacker News new | ask | show | jobs
by slashdotdash 2535 days ago
I’ve personally suffered the same fate as Relive with Strava cutting off access to their API for a side project of mine [1]. They gave me twenty minutes notice via email before disabling access and effectively killing my app.

Ostensibly the reason was due to competing with a feature that Strava intend to build themselves (“thanks for proving the demand”). This was back in March, I’ve since been informed that access won’t ever be reinstated so have open sourced the code [2].

Lesson learnt, don’t depend entirely on third party APIs.

[1] https://segmentchallenge.com/

[2] https://github.com/slashdotdash/segment-challenge

2 comments

I think it's a bit more nuanced than just "don't depend entirely on third party APIs," rather make sure that your API provider benefits from your use of their services expanding. [1] Issues only arise when your success comes at the expense of the API provider, whether thats by poaching a potential audience or otherwise.

[1] https://mkwng.substack.com/p/whose-problem-is-this

In over 3 decades of software development, every single app that I wrote which depends exclusively on a third party API or data set has eventually (and suddenly) died a horrible death when the vendor decided to (a) shut down external access (b) change their API significantly without warning, or (c - more commonly) build it within their app themselves after seeing the demand and traction my 3rd party apps were getting.

I still develop optional 3rd party add ons for other vendors, but I will never ever again rely on a single vendor 'lock in' with any of my apps.

And because you can never know what your API provider wants to expand into next, you effectively shouldn't rely on third party APIs. Strava could very well see that some product using their API is doing well and decide to fill the same niche themselves. Once Strava's product is up and running they can just kill off access to the other one.
Exactly. Just look at the numerous twitter apps that are no longer around but we’re very useful to twitter at the time.
Is providing the service to the API providers competitors not somewhat creating a bit of defense in this situation? If every app can do it through a partner which is focused on that feature, is there a real benefit to the app building their own solution?

I think this is Relive's misstep, though time will tell. They have all of the major apps as part of their API now, but they've also started directly competing with these app, giving the app creators an incentive to compete right back.

I do have a dog in this fight as the founder of https://ayvri.com which creates interactive 3D videos, used for sports and other domains.

It gives the API provider the opportunity to let others take the risk and build a market for a potential feature. Then if the market is there they just cut access to the API for the others and and build that feature into their own app. This way the API provider always reaps the benefits and pushes back any app before they start to compete directly.

And while less developers will be incentivized to use the API seeing this behavior, very few can actually afford to launch a proper competitor. So plenty will continue do develop with this philosophy in mind, knowing that if it's successful they will be cut off but hoping they make enough and then just pivot to the next thing.

It does give them the right to exclude them. The problem is that others will see that and be less willing to write software based on your API. If this is a pervasive thing in the industry then everyone will only want to roll their own services.
> Lesson learnt, don’t depend entirely on third party APIs.

What's funny is that Relive doesn't have to depend on Strava. When I knew about the service months ago I signed up expecting to be able to upload a GPX file and get a video back. But no, they want to get those files via Strava (or other cloud services) so I closed my account.

I understand that 1) it removes a lot of friction, 2) it's what the users of Strava are used to and 3) they can get all those juicy data instead of a few files. However it's risky and Strava eventually cut them off.

I wonder if they'll implement a file upload, just live with the other services, or close the business.

Relive actually does support GPX upload and touts it as a way to still integrate with Strava, albeit by manually downloading the gpx from strava and uploading to relive: https://support.relive.cc/article/18-import-old-activity
Good to know because I don't want to have an account on Strava or similar services. I didn't notice that import functionality when I created my account, only the prompts to connect to third-party services. I'll create a new account and check it out.