5 ms·

I’ve been subscribed for several years. It’s excellent. The general approach inspired something I’m working on now on the side.

mikeryan 5y ago
Awesome. I’d offer any advice but technically this was pretty straightforward. It’s React, Auth0 (see note), Stripe and Contentful with a few Node services to glue the SaaS pieces together. * So originally we had social auth. We didn’t know how, or if, this would be monetized, but we knew we wanted to force sign up so we used Auth0. When we launched we told to expect 10,000 (maybe) MAU. The weekend we launched we got 250k signed up users. It was all cloud services and we weren’t worried about scale, our one freak out was when we were blowing through our 10k users tier with Auth0 into Enterprise “call us” level on a weekend. Auth0 just let it ride and we dealt with it later. But Neil passionately despises Facebook and Google so he ended up not wanting to allow social log ins. When he found out 2/3 of our users were on one of those two endpoints he relented. (Of note it was pretty much 1/3 FB, 1/3 Google, 1/3 User/Pass) He has since removed all Social providers.
- 52-6F-62 5y ago
  Well we’re aligned on the social media front. I’m working on spinning up a label for managing a personal release, and as part of it want to be able to self host and stream some tracks from releases or singles for free without just linking an easily interceptable mp3 in an audio element. So, mainly inspired from his efforts with Orastream, but of course without infrastructure there. To save myself rambling further, I’m aiming to release, open sourced under the label, a containerized boilerplate for self hosting a server and client nodes that operate seekable audio streaming (for FLAC at first) and standard artist site content (bio, media, etc). Mainly as an option for artists with a bit of know how but not enough resources for something more robust so that they can operate independently of any other services if they wanted. From there I’m hoping people would see enough value to contribute and to build on it from there. I’ll definitely take your cues. While I doubt I’ll struggle any time in my life with Neil’s kind of numbers, maybe someone else would. Thanks for the insights! That guy has been a hero of mine since I was young. Still is.
  - mikeryan 5y ago
    Yeah OraStream really isn’t that magical it’s based on a spec (I can’t find it right now) that’s basically just a variation of HLS. Audio only HLS with high bitrate renditions with pretty much any encoder (ffmpeg) should work as a pretty straightforward drop in for what OraStream does. Dash should work as well.
    - 52-6F-62 5y ago
      That's more or less the plan! But on a small scale for self-deployment. Thankfully had a little experience with high bitrate streams at Rogers, at least superficially. Did your team build the player itself as well? The meter for bitrate was a nice touch!

mikeryan 5y ago

Awesome. I’d offer any advice but technically this was pretty straightforward. It’s React, Auth0 (see note), Stripe and Contentful with a few Node services to glue the SaaS pieces together. * So originally we had social auth. We didn’t know how, or if, this would be monetized, but we knew we wanted to force sign up so we used Auth0. When we launched we told to expect 10,000 (maybe) MAU. The weekend we launched we got 250k signed up users. It was all cloud services and we weren’t worried about scale, our one freak out was when we were blowing through our 10k users tier with Auth0 into Enterprise “call us” level on a weekend. Auth0 just let it ride and we dealt with it later. But Neil passionately despises Facebook and Google so he ended up not wanting to allow social log ins. When he found out 2/3 of our users were on one of those two endpoints he relented. (Of note it was pretty much 1/3 FB, 1/3 Google, 1/3 User/Pass) He has since removed all Social providers.

52-6F-62 5y ago
Well we’re aligned on the social media front. I’m working on spinning up a label for managing a personal release, and as part of it want to be able to self host and stream some tracks from releases or singles for free without just linking an easily interceptable mp3 in an audio element. So, mainly inspired from his efforts with Orastream, but of course without infrastructure there. To save myself rambling further, I’m aiming to release, open sourced under the label, a containerized boilerplate for self hosting a server and client nodes that operate seekable audio streaming (for FLAC at first) and standard artist site content (bio, media, etc). Mainly as an option for artists with a bit of know how but not enough resources for something more robust so that they can operate independently of any other services if they wanted. From there I’m hoping people would see enough value to contribute and to build on it from there. I’ll definitely take your cues. While I doubt I’ll struggle any time in my life with Neil’s kind of numbers, maybe someone else would. Thanks for the insights! That guy has been a hero of mine since I was young. Still is.
- mikeryan 5y ago
  Yeah OraStream really isn’t that magical it’s based on a spec (I can’t find it right now) that’s basically just a variation of HLS. Audio only HLS with high bitrate renditions with pretty much any encoder (ffmpeg) should work as a pretty straightforward drop in for what OraStream does. Dash should work as well.
  - 52-6F-62 5y ago
    That's more or less the plan! But on a small scale for self-deployment. Thankfully had a little experience with high bitrate streams at Rogers, at least superficially. Did your team build the player itself as well? The meter for bitrate was a nice touch!