memo: private, keyboard-first notes for your terminal

Why memo

Capture a thought without leaving your flow

Mid-debug idea, standup notes, a TODO from a code review: jot it in the next pane over. Everything runs on your machine: the editor, the AI, the speech model.

🔒

Private by architecture, not by policy

There is no server to trust, no account to create, no telemetry to opt out of. Refinement and dictation run on your hardware, and your notes are files in your home directory. Nothing ever leaves your machine, because there's nowhere for it to go.

Keyboard-first TUI

Press n, type, done. Vim-adjacent muscle memory, folders, search, and export; no mouse required, no Electron in sight.

✦

Local AI refinement

Ctrl+R cleans up grammar and adds Markdown structure using a model on your machine (MLX or Ollama). The original is always kept, side by side.

🎙

Offline dictation

Hold F5 for push-to-talk, double-press for continuous listening. Speech-to-text runs locally via whisper.cpp: nothing leaves your device.

.md

Plain Markdown, no lock-in

YAML frontmatter, atomic saves, fully portable. grep them, pipe them, commit them to a repo: your notes are readable in any editor, forever.

▣

Lives in your workflow

It's just a TUI in a shell: drop it in a tmux pane, a zellij tab, an SSH session, or Warp. If your setup renders a terminal, memo already works there.

🦀

One Rust binary

No daemon, no background sync, no runtime to install. A single fast binary for macOS, Linux, and Windows that starts before you finish the thought.

A command palette, one keystroke away

Type / at the start of a line, or Ctrl+O from anywhere in it, and search as you type. Drop in a to-do, callout, link, heading, or hidden comment, or fire an action like Change title and Refine note. Ctrl+L ticks a to-do off without reaching for the palette at all.

Install

Up and running in 30 seconds

One command installs memo and sets up a local AI backend for you. Read the script first if you like; it's open source, like everything else here.

$ curl -fsSL https://blackstardesigns.github.io/memo/install.sh | sh

On Apple Silicon the installer sets up MLX; on Intel Macs and Linux it uses Ollama. Prefer a manual install? Grab a prebuilt binary.

$ powershell -c "irm https://blackstardesigns.github.io/memo/install.ps1 | iex"

$ git clone https://github.com/blackstardesigns/memo.git $ cd memo $ cargo run

Requires Rust, CMake, and a C/C++ compiler: whisper-rs compiles whisper.cpp at build time. See the docs.

Notes for developers who live in the terminal, refined by local AI.