Alexa Skill Stream Not Working? Here Are the Real Fixes (HTTP vs HTTPS and More)

"It plays in VLC but not on Alexa" is the most common support request in this entire industry, and it's almost never mysterious. There are six causes, they account for nearly every case, and five of them you can check yourself in under ten minutes. The single biggest is an HTTP stream URL — the other five are the long tail, but each has a distinctive signature once you know what to look for.

Cause 1: HTTP Instead of HTTPS

Amazon requires HTTPS, full stop. Desktop players and browsers are lax — they'll happily play an insecure stream — so a station that "works everywhere" often means it works everywhere that doesn't enforce security. The Echo enforces it. Look at the URL: no 's' after http, that's the bug, that's the whole bug.

1. 1On Zeno.FM, Radio.co, or Live365: your URL is already HTTPS — this isn't your problem, move to cause 2
2. 2On managed Shoutcast/Icecast hosting: ask your host for the SSL-enabled port or proxy URL — most provide it free, same day
3. 3Self-hosted: put a reverse proxy in front of the stream — Caddy with Let's Encrypt takes about 20 minutes and renewals are automatic
4. 4After the change: update the stream URL inside your skill — the skill is still pointing at the old URL until someone changes it

Cause 2: The OGG Problem

VLC plays every codec ever devised. Alexa plays three: MP3, AAC, HLS. A station streaming OGG Vorbis or Opus — the default on many Icecast and AzuraCast setups — plays flawlessly in VLC and is pure silence on an Echo. This combination (works in VLC, dead on Alexa) is the classic OGG signature.

The fix is adding an MP3 or AAC mount point on the server and pointing the skill at that. Keep the OGG mount for web listeners if you like it — just don't send it to Alexa.

Cause 3: Port Blocking

Classic stream URLs carry a port number: https://stream.example.com:8000/live. Plenty of networks — corporate firewalls, some ISPs, some enterprise Wi-Fi — filter traffic on non-standard ports, and some of the paths Alexa's audio requests travel through do too. A stream that works at home and fails at the office is usually this.

The robust fix is moving the stream to standard port 443 via a reverse proxy, so the URL looks like an ordinary website: https://stream.example.com/live. Almost every professional setup ends up here eventually, because it fixes causes 3 and 4 in one move.

Cause 4: Mobile Carrier Filtering

"Works on Wi-Fi, dies on mobile data" — some carriers, particularly in the Middle East, Africa, and parts of Asia, throttle or block audio on unusual ports and non-standard protocols. Test the skill or stream on cellular, not just Wi-Fi, before declaring anything broken. Same fix as cause 3: a 443-based HTTPS stream is indistinguishable from normal web traffic and sails through.

Cause 5: The Stream Was Down During Certification

Amazon's testers play the stream at intervals across 1–2 days. A station that goes dark overnight — encoder crash, AutoDJ schedule gap, provider maintenance window — fails for unavailability even though it was up at submission time. Before resubmitting: encoder up 24 hours straight, AutoDJ fallback active, no maintenance scheduled. If the stream has a dropout pattern, Alexa listeners hit the same dropouts — fix the pattern, not just the submission.

Cause 6: A Stale Endpoint After Moving Hosts

After a host migration, old URLs do one of three things: redirect, time out, or fail silently. Browsers follow the redirect, so "it sort of works" in a browser. Alexa handles redirects poorly, so it doesn't. If you've switched providers recently and the skill broke, the skill is still pointing at the old URL — it needs the new direct URL, not a redirect.

The Ten-Minute Diagnostic

1. 1Play the stream in VLC on desktop — establishes the stream itself is up
2. 2Read the URL — https? No 's' = cause 1, fix that first
3. 3VLC → Tools → Codec Information — MP3 or AAC? OGG = cause 2
4. 4Play the stream over mobile data, Wi-Fi off — carrier filtering reveals itself here (cause 4)
5. 5Has the stream been up 24+ hours without a drop? If not, cause 5
6. 6Moved hosts in the last few months? Cause 6 — update the skill's endpoint

That sequence finds nearly every failure we've ever diagnosed. And notice what isn't in it: nothing about the skill's code, Amazon's infrastructure, or anything you can't see from your own desk. Alexa stream failures look intimidating from the outside and are almost always one of these six.