LoFi Engine generates its tracks procedurally with Tone.js, so there is no file to loop

The music is synthesised, not streamed or bundled

The core technical claim of the project is in one line. The lofi track is generated procedurally using Tone.js, providing a dynamic and unique listening experience every time.

That is the whole product in one sentence, and it explains the offline claim as a consequence rather than a feature. If the music is computed on the device, there is no server to stream from, no catalogue to sync and no network requirement, and the README states plainly that it does not require an internet connection and that the experience is fully private.

Tone.js is a web audio framework for building synthesised instruments and effects in the browser, and it is the only audio dependency in the package manifest, at version 14.9.17. Everything else in the dependency list is interface or shell related.

The result is that the application is a sequencer and a set of instruments rather than a player. The naming reflects that. The repository is called lofi-engine, the package is `LofiEngine`, and the release tags are `app-v1.0.3`, `app-v1.2.0` and `app-v2.0.0`, so the project distinguishes the application from the engine inside it.

The consequence for a user is that there is no track list in the usual sense. The music is something you shape rather than something you choose, and the README makes that explicit with the phrase Be the Artist in the customisation section.

Auto DJ takes over the mixing, and its four modes are the real interface

The Advanced Immersion system, code name `Auto DJ`, automatically manages the soundscape for you, and it offers four modes.

Music focuses on the beat and chords without atmospheric effects, so it is the plainest option. Atmosphere adds core weather and nature effects for a fuller sound. World incorporates specific textures such as city sounds or wind for deep immersion. Manual gives you full control to mix and match sounds as you please.

That progression is sensible. Music is the lofi skeleton, Atmosphere layers weather and nature on top of it, World pushes further into a specific place, and Manual hands the whole mix back to you. The naming is what makes it clear that these are levels of automation rather than four different playlists.

The customisation section frames the same idea in three axes. Playback is where you decide whether you want music or, in the README's phrasing, just birds singing freely. Look is about the artwork, with carefully crafted artworks from talented artists, aimed at making long study sessions bearable. Ambience covers the sound effects, from a windy evening to the sound of waves crashing into the sand of a sunny beach.

The stated main focus of the project is customization, accessibility and artistic freedom. Those three words describe the design priority honestly, and they are worth reading as the answer to the question of what this application is for: it is a configurable generative music environment rather than a listening app.

The productivity layer is a Pomodoro timer with statistics

LoFi Engine ships with what the README calls productivity features, and they are the part that separates it from a generic generative music app.

There is a Pomodoro timer with configurable focus and break durations, automatic phase switching and desktop notifications. There are focus tasks, so you can select what you are working on, track completed Pomodoros and keep completed work organised. There are focus stats with daily and weekly totals plus a seven-day focus-time graph. There is a Zen mode that hides the track list and effects controls while the scene and music continue. And there are desktop conveniences, namely a mini player mode and a system-tray menu covering window visibility, play and pause, phase skipping and quitting.

The system tray integration is the detail that makes it a desktop app rather than a web page. Being able to start a Pomodoro from the tray and have the notification appear when the phase changes means the window does not have to be in front of you, which is the entire premise of a focus tool.

The desktop notifications come from a Tauri plugin, `@tauri-apps/plugin-notification` at 2.4.0, which confirms the approach. The other Tauri plugins in the manifest are for the shell, for the window state, and the API package itself.

The focus tasks and stats are the part that gives the timer a reason to exist beyond the browser tab timers people already have. A pomodoro counter in a music app is a gimmick, and a pomodoro counter that logs what you worked on and shows you a week of focus time is a tool. Whether the tracking is worth the friction is a matter of taste, and the focus tasks being optional rather than required suggests the author thought about that.

Seven languages, and every mouse action has a keyboard shortcut

Two accessibility commitments are stated, and both are concrete rather than aspirational.

The first is internationalisation, with support for English, French, Spanish, Japanese, Korean, Indonesian and Russian. The app locale is a real concern for a desktop app with a fixed menu bar, and seven languages including Korean and Japanese is a commitment of translation work rather than a flag.

The second is keyboard access. Any action or click you can do with the mouse has a shortcut for it with the keyboard, for playback, effects, ambient tracks and look among others. All information about the app and the shortcuts is available in an info box accessible via the Right Click context menu, under About.

That last detail is the one that shows the feature was designed rather than bolted on. A shortcut list that is only in the documentation is useless once you have forgotten it, so putting it in the app itself behind a context menu means the help is always one right click away, and that the same gesture that opens the info box is the gesture that opens the context menus in the first place.

Tauri for the shell, Svelte for the interface, and a Rust build requirement

The tech stack is Svelte with TypeScript, Vite as the bundler, Tauri for the desktop shell, pnpm as the package manager, and Tone.js for the audio. The repository layout follows that split, with a `src/` directory for the Svelte application, a `src-tauri/` directory for the Rust side of the Tauri shell, a `public/` directory for static assets and a `scripts/` directory.

Tauri rather than Electron is a deliberate choice with a visible consequence. The app ships a web view provided by the operating system instead of bundling Chromium, so the installed footprint is much smaller. The cost is that you need the platform's native web view to be present and to behave the way you expect, and that you need a working Rust toolchain to build at all.

The prerequisites section lists Node.js v14 or later, pnpm v6 or later, Rust at the latest stable version, and the Tauri prerequisites for your operating system. So a full local build needs Node, pnpm, Rust and a native toolchain, which is a heavier setup than a typical JavaScript project. Once the toolchain is in place, the install and the development run are two commands.

pnpm install
pnpm tauri:d

The first installs the dependencies, and the second starts the Vite dev server for the frontend together with the Tauri development process for the native shell. The production equivalent is `pnpm tauri:b`, which writes a release build into the `src-tauri/target/release` directory.

The scripts in the package manifest are short and mirror that split. `dev` and `build` are plain Vite, `preview` serves the built frontend, `check` runs svelte-check against the tsconfig, and the three Tauri scripts are `tauri`, `tauri:d` for development and `tauri:b` for a production build. The Vite-only commands are the useful ones for a frontend contributor, since they let you work on the interface without installing Rust.

There is also a `pnpm-workspace.yaml` at the root, so the project is set up as a workspace, and both a `pnpm-lock.yaml` and a `package-lock.json` are present.

The release history shows a 2.0.0 in September 2026 after 1.2.0 in December 2025 and 1.0.3 in November 2025, and the last push to the repository was on 2026-09-22. The project is MIT licensed, and it is deployed at lofi-engine.vercel.app, which is the web build, with the desktop releases distributed through GitHub releases.

Editorial conclusion

Adopt LoFi Engine if you want a focus soundtrack that is generated on your machine, never repeats exactly and needs no network, and you are on Linux, macOS or Windows. Do not adopt it if you want a curated library of finished tracks with known licences, because there is no catalogue here at all, only synthesis. Verify first that the Tauri prerequisites on your platform are installed, because the build needs Rust and a native toolchain, and that the audio output path works on your machine, since a synthesis app that has to be routed through a specific interface is a different proposition from one that plays to speakers.

Frequently asked questions

Official sources

1. License: MIT
2. meel-hd/lofi-engine on GitHub
3. Project website
4. README
5. Releases