Thingmo

The Problem

Streaming recommendation engines optimize for engagement. They are good at keeping you listening to things you already like. They are not good at helping you find music that connects to something specific — an artist's stated influences, a moment in a city's music history, a scene you read about but never heard.

Thingmo is built for a different kind of curiosity.

How It Works

Instead of recommending music based on behavior, Thingmo builds playlists from real-world inputs: artists and the records they explicitly pointed to, historical charts tied to specific dates and locations, and scenes defined by time and place.

Each search produces a clean, playable playlist that can be exported to services like Apple Music and Spotify.

Design Principles

• No accounts, no feeds, no social features — this is a tool, not a platform
• Results grounded in documented history, not behavioral inference
• Fast path from idea to something you can hear immediately
• No algorithmic drift — the same search returns the same kind of result
• Export-first: the goal is music in your player, not time on the site

Why It Exists

Music discovery tools are mostly built around what the algorithm thinks you want next. Thingmo is built around what you're actually curious about. The difference is whether the tool follows your intent or substitutes its own.

That distinction — user intent versus system intent — is the same tension at the center of most AI product design problems. Thingmo is one place I work through it in a concrete domain.

Current Status

In development. Core search model and playlist generation are functional. Export integration and UI are in progress.