the short version

you type a moment in plain language. it can be a scene (“morning kitchen after last night”), an activity (“music for fixing a car at 2am”), a feeling (“antistress”, “trening”), or anything else specific to where you are right now. the music starts. it keeps going. when you want something different, you type a new moment and the stream shifts.

no library to browse. no playlist to maintain. no songs you've already heard three times this week.

why this shape

an early user wrote me one sentence that explains the whole product: “spotify or youtube music compilations get a bit boring after a while. i mostly listen to music in the background while doing other things. if your service continues to work, that would be great.”

that's the wedge. most of the music you listen to in a day isn't the kind you sit with and study. it's the kind that's under the rest of your life: coding, cooking, driving, getting ready, decompressing, falling asleep. and the way you currently get that music is fine until you've been doing it for a few weeks and the playlists start to loop.

the closest existing reference for the shape i wanted is the lofigirl youtube livestream. one stream, no ending, you put it on and forget it. the limitation is that it's only lofi. flowy is the same shape, but you describe the music.

not a song generator

flowy is sometimes confused with suno or udio. they're both ai music, but the jobs are different. suno makes one song at a time for creators to iterate on. flowy makes a stream for listeners. there's no save button on a track in flowy, no “extend by 30 seconds”, no lyric editor. the whole product is the stream and your description of where you are.

i compare against brain.fm and spotify's focus playlists instead. those are listener tools too. flowy is the bet that “describe a moment” beats “pick a preset” (brain.fm) or “pick a playlist” (spotify).

who's building this

tim grossmann. multi-time founder and hands-on ai product builder, based in germany.

the path here, short version: instapy (open-source automation that picked up enough users to matter), explo (a data product i ran through real customer discovery and a real growth curve), with stretches at bosch and volkswagen on the engineering side along the way. that mix is the reason i'm equally comfortable shipping code at 2am and writing the memo a buyer at a bigger company needs before they say yes.

today my center of gravity is blockbrain and tigr ventures. blockbrain is where i work with smbs and internal champions on small, practical ai pilots and workflow automation, the kind of work where the person you're talking to can actually decide to ship something this quarter. tigr ventures is the operating entity for flowy and a few other useful-ai side products.

if i had to compress what i do: builder, operator, connector. strong bias for fast experimentation and useful ai over abstract ai. flowy is the slice of that pointed at my own listening habits.

the german legal stack on flowy is real (imprint, agb, widerruf, dsgvo, all compliant from day one) because if i'm going to run a paid product, i'd rather build it under the strictest consumer-protection rules and have the rest of the world come for free.

i read every email that comes in. if something is broken, confusing, or just feels off: contact.timgrossmann@gmail.com.

what flowy is good at right now

specific scenarios. the more detail in your moment, the better the stream. weather plus time plus activity is a great combination.
retuning mid-stream. switching from one moment to another is smooth. you don't restart.
regional music. moments in other languages (corridos, perreo, bollywood, k-pop, japanese city pop) produce music with vocals in the matching language and the right regional sonic vocabulary.
no repeats. tracks don't loop within a session. signed-in users get the stronger guarantee: tracks you've heard in any prior session never replay.

what's still mid

in the spirit of not lying:

pure-genre prompts drift. a russian dj sent me feedback last week that typing “afrohouse” sometimes returned tracks that weren't afrohouse. he's right. the fix is to write scenarios rather than genres alone (“afrohouse warmup for guests arriving” lands more reliably), but the underlying model fidelity needs work and i'm on it.
playback occasionally stops. a small number of early users have had to reload the page mid- stream. that's the single thing i care about fixing first. if it happens to you, please email me with what you were doing right before it broke.
russian payments are still a problem. roughly a third of early traffic is from russia, and stripe doesn't accept russian cards. the workaround right now is a foreign card or crypto. native crypto checkout is the next big piece of work.

pricing

you can listen without an account: three plays per station, two stations per browser. sign in for free and you get 90 minutes a day, saved tracks, and full retune support. unlimited streaming and every download free is on the subscription tier, which sits below the price of comparable services on purpose. 14-day refund window on anything you pay for, per eu consumer law.

if it works for you, it should be cheaper than alternatives. if it doesn't work for you, you should be able to walk away clean.

what's next

playback reliability. the single most important thing.
genre fidelity. pure-genre prompts should hit harder than they do today.
crypto checkout. for the russian listener segment specifically.
library hub. a public directory of the seeded moments with real audio. ~300 pages of curated vibes browsable on the open web.
shareable sessions. every stream you start could have a public url. opt-in.

none of this is a startup pitch deck. it's a small tool i wanted for myself, and it turns out other people wanted it too. if you're one of them, type a moment and see if the stream feels right.