Install
Run fontego-server in Docker or from source, and set up the first admin and workspace without questions.
Fontego is one server with its data in one folder: every workspace (a Git repository each) and an SQLite database with the accounts and the logs. It needs git 2.38 or newer, which the container already has.
In Docker
From a checkout of fontego-app with fontego-cli next to it (the image takes @fontego/core from there):
docker compose up -d --waitcompose.yaml runs the container locked down: it writes only to its data volume and to /tmp, holds no Linux
capabilities, can't gain privileges, and has a ceiling of memory, CPU and processes, so whatever happens inside stays
inside. Never mount the Docker socket into it.
These variables, in the environment or in a .env file next to compose.yaml, change how it runs:
| Variable | Default | |
|---|---|---|
FONTEGO_BIND | 127.0.0.1:8740 | Where the port is published. Keep it on 127.0.0.1 and put a proxy with HTTPS in front |
FONTEGO_PUBLIC_URL | http://localhost:8740 | The address people and their CLIs use |
FONTEGO_MEMORY | 1g | Memory limit; raise it for big workspaces |
FONTEGO_CPUS | 1.0 | CPU limit |
Mail, a config file and backups go in its environment section; compose.yaml has each one commented. Without
Compose:
docker build --build-context fontego-cli=../fontego-cli -t fontego-server .From source
With Bun 1.3.14 or newer and git 2.38 or newer:
git clone https://github.com/fontego-ai/fontego-app.git
git clone https://github.com/fontego-ai/fontego-cli.git
cd fontego-app && bun install
FONTEGO_DATA_DIR=/var/lib/fontego bun server/src/main.ts serveThe data folder is created with mode 0700: only the server's user may enter it.
The first run
fontego-server init sets up a new instance without asking anything, so a script or an agent can install Fontego
from start to finish:
fontego-server init you@example.org --display-name "Our team" --workspace notes --template demo --token- The server admin: the email given, with a link to choose a password, which
initprints. - A token for the CLI, with
--token, printed once. - The visible name of the web app, with
--display-name. An admin can change it later. - The first workspace, with
--workspace SLUG, and the admin as its admin. It starts empty, or filled from:--template NAME: a template.fontego-server templateslists them; the server ships the ones inexamples/, and the config'stemplatesDiradds your own, one folder each.--from PATH: a folder or a.zipon the server's machine. In Docker, mount it first (the/importvolume incompose.yaml).--git URL|PATH: a Git repository, with its history;--branchpicks the branch and--no-historytakes only its last version.
Each option has its key in the [init] table of the config file, so one file can describe a whole install; options on
the command line win. Running init again changes nothing that is already there.
Every file brought in goes through the checks of any write (secrets, size, rejected paths) and the ignore rules the
source carries (its .fontegoignore files). What is left out is listed. A failed import leaves no workspace behind.
In Docker, run it inside the container:
docker compose exec server fontego-server init you@example.org --workspace notes --template demo --tokenThen
- Sign in at the link
initprinted, or with the CLI:fontego login <public URL>. - Invite people: People and accounts.
- Before others use it: Production.
Every command of fontego-server is in the server reference.