Skip to content

INSTALL.md

Latest commit

 

History

History
68 lines (49 loc) · 2.48 KB

INSTALL.md

File metadata and controls

68 lines (49 loc) · 2.48 KB
Raw

Installing Zedium

Zedium is a telemetry-free, cloud-free fork of Zed. It ships as a plain archive — there is no auto-updater and no account.

Platforms

Platform Artifact
Linux x86_64 zedium-<tag>-linux-x86_64.tar.gz
Linux aarch64 zedium-<tag>-linux-aarch64.tar.gz
macOS arm64 zedium-<tag>-macos-arm64.zip (contains Zedium.app)
macOS x86_64 zedium-<tag>-macos-x86_64.zip (contains Zedium.app)

Tags use the form v1.4.2-1 (upstream tag + fork revision).

Linux

tar xzf zedium-<tag>-linux-x86_64.tar.gz
cd zedium-<tag>
./bin/zedium

To put it on your PATH, symlink bin/zedium into ~/.local/bin/.

User data lives under XDG paths keyed to Zedium, so it never collides with an upstream Zed install:

  • Config: $XDG_CONFIG_HOME/zedium (default ~/.config/zedium)
  • Data: $XDG_DATA_HOME/zedium (default ~/.local/share/zedium)

macOS

The app is ad-hoc signed (no Apple Developer ID). Gatekeeper will quarantine it on first launch. Remove the quarantine attribute:

unzip zedium-<tag>-macos-arm64.zip
xattr -dr com.apple.quarantine Zedium.app   # -r: clear the whole bundle
open Zedium.app                              # or drag Zedium.app into /Applications

Bundle identifier: dev.zedium.Zedium. User data: ~/Library/Application Support/Zedium, config ~/.config/zedium.

Remote development (SSH / Docker / WSL)

Remoting works offline: the pre-compiled remote_server daemons for Linux (x86_64 + aarch64) and macOS (arm64) ship inside the app, next to the editor binary (Zedium.app/Contents/MacOS/remote_servers/ on macOS, bin/remote_servers/ on Linux). At connect time the editor provisions the matching daemon onto the host — no download from the network. See REMOTE_SERVERS.md.

Updating

There is no in-app updater. Download the newer archive and replace the binary/app bundle. Your config and data directories are preserved across versions.

First run

  • No sign-in, no telemetry, no network calls on startup.
  • AI features (assistant + edit prediction) are off until you configure a provider. See EDIT_PREDICTION.md for self-hosted edit prediction; configure chat models under language_models in settings.json (Ollama / LM Studio / any OpenAI-compatible endpoint).
  • The extension registry is a static offline snapshot bundled with the release; there are no live extension updates between releases.