Tunio

About a year and a half ago, I wanted something pretty simple:

a radio station I would actually listen to.

I wanted tech news, gaming, movies, weather, podcasts and music — but without constantly switching between apps and sources.

So, being a developer, instead of finding a reasonable solution, I started building one. :)

The first version was basically a pile of scripts.

RSS feeds were collected and summarized, TTS turned them into short radio news segments, weather was generated automatically, and AI-generated music filled the gaps.

Then I added jingles.

Then scheduling.

Then Icecast.

At some point it became a 24/7 radio station that could basically run without me.

And honestly, that was supposed to be the end of the project.

Then I made the mistake of showing it to other people

I built a simple UI and shared it with a few friends.

The interesting part was that they didn’t necessarily want to build radio stations.

One wanted background music for a store.

Another wanted it for a gym.

They cared about completely different things:

• Can I schedule different music for morning and evening?

• Can it keep playing if the Internet goes down?

• Can I insert announcements between tracks?

• Can I run it on a TV or cheap Android box?

• Can I use music without worrying about commercial licensing?

That was the moment I realized I might have accidentally built the wrong product for the right problem.

The radio wasn’t really the product.

The automation behind the radio was.

So I started turning the hack into Tunio

The idea became: what if running audio for a business was basically programmable?

You define what should happen, and the system keeps the stream running.

Music at 9 AM.

An announcement at 11.

Different playlist after lunch.

Weather or generated content when needed.

A live broadcast occasionally.

And everything continues 24/7 without someone sitting there managing playlists.

Technically, this rabbit hole went much deeper than I expected.

I originally used Liquidsoap + Icecast, but eventually wrote my own lightweight broadcaster in Go around FFmpeg.

I discovered that cheap VPS instances and real-time audio are not particularly good friends.

CPU contention caused random audio artifacts, so eventually I moved the streaming workload to dedicated servers.

I also made the mistake of building around MP3 first, only to discover later that restreaming to YouTube/Twitch made AAC a much better foundation.

There were many moments like this.

The funny thing is that I originally started the project partly because I wanted an excuse to experiment with distributed systems and broadcasting infrastructure.

Be careful what you wish for.

AI changed the project more than I expected

Initially I thought AI would mostly help generate content.

Instead it ended up affecting almost every layer of the product.

Music can be generated.

Announcements and jingles can be generated.

Voices can be cloned.

RSS articles can become short audio news segments.

Podcasts can be generated.

And lately I’ve also been using coding agents for boring product work — localization, documentation, screenshots and repetitive engineering tasks.

So instead of building another streaming service with AI features attached to it, I’m increasingly thinking about Tunio as an audio system where both the programming and the content can be generated dynamically.

That’s a much more interesting problem to me.

Where it is now

What started as my personal radio hack is now an open beta called Tunio.

People are using it for online radio, stores, gyms and other places that need continuous background audio.

I’m still figuring out exactly where the strongest market is.

And that’s actually why I’m posting this here instead of writing another polished launch announcement.

I’m curious how other founders would think about the positioning.

If you landed on this product today, which description would make more sense to you?

1. AI radio platform

2. Background music platform for businesses

3. Programmable audio platform

The third one is closest to how I personally think about what we’re building — but I’m not sure it’s the clearest one for someone seeing Tunio for the first time.

Would love to hear what you think.