Skip to content

Installation

Kody Dennon edited this page Jul 4, 2026 · 3 revisions

Installation

This page lists every supported way to install ConvexAutoBackup. For most people, Docker or the native installer is the easiest path. Cargo is best for Rust users who already manage their own services.

Requirements

Path Required
Native installer macOS, Linux, or Windows with network access to GitHub Releases.
Docker Docker Engine or Docker Desktop.
Cargo Rust stable toolchain, plus your own service/process setup.
Source Rust stable, Node.js, npm, git, and make.

You also need a Convex deploy key for real backups.

Native beta installer

macOS and Linux:

curl -fsSL https://github.com/KodyDennon/ConvexAutoBackup/releases/download/v0.1.0-beta.5/install.sh | sh

Windows PowerShell:

iwr https://github.com/KodyDennon/ConvexAutoBackup/releases/download/v0.1.0-beta.5/install.ps1 -OutFile install.ps1
powershell -ExecutionPolicy Bypass -File .\install.ps1

The native installers are intended to be the normal path for people who want a system-managed app. They download the right release bundle, verify checksums, generate a CONVEX_AUTOBACKUP_MASTER_KEY, provision the pinned Convex CLI runner, and configure autostart by default.

Opt out of autostart:

curl -fsSL https://github.com/KodyDennon/ConvexAutoBackup/releases/download/v0.1.0-beta.5/install.sh | sh -s -- --no-autostart

Windows:

powershell -ExecutionPolicy Bypass -File .\install.ps1 -NoAutostart

Open the web UI:

http://localhost:8976

If the install is on another LAN/server host, use that host name or IP address with port 8976.

Windows MSI

Windows releases include an MSI for users who prefer a package installer. The MSI installs binaries. The PowerShell installer remains the fuller setup path because it can also provision the runner, write generated config, and configure a service.

Docker install script

curl -fsSL https://github.com/KodyDennon/ConvexAutoBackup/releases/download/v0.1.0-beta.5/docker-setup.sh | sh

The script creates a persistent install directory, generates a master key, starts the supervised service, and checks health.

Manual Docker

Docker Hub:

docker pull kodydoty/convex-autobackup:latest
docker run -d \
  --name convex-autobackup \
  --restart unless-stopped \
  -p 8976:8976 \
  -v convex-autobackup-data:/data \
  -e CONVEX_AUTOBACKUP_MASTER_KEY="$(openssl rand -base64 32)" \
  kodydoty/convex-autobackup:latest

GHCR:

docker pull ghcr.io/kodydennon/convex-autobackup:latest
docker run -d \
  --name convex-autobackup \
  --restart unless-stopped \
  -p 8976:8976 \
  -v convex-autobackup-data:/data \
  -e CONVEX_AUTOBACKUP_MASTER_KEY="$(openssl rand -base64 32)" \
  ghcr.io/kodydennon/convex-autobackup:latest

Open:

http://localhost:8976

Docker Compose from source

git clone https://github.com/KodyDennon/ConvexAutoBackup.git
cd ConvexAutoBackup
CONVEX_AUTOBACKUP_MASTER_KEY="$(openssl rand -base64 32)" docker compose up --build

The Compose setup exposes port 8976 and persists /data.

Cargo install

cargo install convex-autobackup --version 0.1.0-beta.5
convex-autobackup runner install --json
convex-autobackup supervise

Cargo installs the CLI/server binary. It does not create OS services, generate managed config, or install Docker volumes. Use Docker or native installers for the lowest-friction setup.

From source

git clone https://github.com/KodyDennon/ConvexAutoBackup.git
cd ConvexAutoBackup
make setup
make check
cargo run -p convex-autobackup -- supervise

Open:

http://localhost:8976

First run

On first run, create the owner account in the web UI or use the bootstrap API. Existing installs show normal sign-in and a local recovery command for creating another owner from the server shell.

After sign-in, the web console guides setup one task at a time:

  1. Create a project.
  2. Store or reference a Convex deploy key.
  3. Create a Convex Cloud target.
  4. Create a local or S3-compatible destination.
  5. Create a backup job.
  6. Create a schedule.
  7. Run and verify the first backup.

Manual configuration remains available in the setup page for non-linear edits and advanced cases.

Clone this wiki locally