Creates a Sillo application from a starter repository.
uvx sillo-start create-app myapp
cd myapp
make setup
make devThat is the whole tool. It fetches a real, working application, renames it to yours, gives it its own secrets, and gets out of the way.
A generator renders templates. Templates are checked for rendering, which is not the same as working — a project can produce valid Python, import cleanly, render every page, and still fail on its first real request. Middleware registered in the wrong order, an auth backend reading the wrong claim, a missing static mount: all of them render perfectly.
sillohq/starter is a real application
with its own CI. Every push boots it and exercises every route, on three
Python versions. What you get has been run, not just written.
It also means the starter can be read, forked and improved on its own, and
that sillo-start never has to be released to fix a bug in what it produces.
# Run it without installing
uvx sillo-start create-app myapp
# Or install it
uv tool install sillo-start
pipx install sillo-startPython 3.11 or newer. The only dependencies are typer and rich.
sillo-start create-app myapp # the default starter
sillo-start create-app sillohq/starter myapp # named explicitly
sillo-start create-app sillohq/starter@v1.2 myapp # pinned to a tag
sillo-start create-app acme/our-template myapp # your own starterAny public GitHub repository works. A full URL is accepted too, so pasting what is in your address bar does what you expect.
| Option | |
|---|---|
--ref <branch|tag> |
Which revision to take. Defaults to main |
-d, --directory <path> |
Where to create it. Defaults to ./<name> |
--install |
Install dependencies straight away |
--no-git |
Do not initialise a git repository |
--force |
Allow a directory that is not empty |
-v, --verbose |
Show tracebacks |
Dependencies are not installed by default, so creating a project takes a
second rather than a minute. make setup in the new project does it, along
with everything else a first run needs.
The starter arrives as a tarball rather than a clone. That needs no git on
the machine, brings no history for you to delete before your first commit, and
pins to a tag as easily as to a branch.
Then three things happen:
It takes your name. The package name, the application title, the SQLite path. Rewriting is targeted rather than a blanket find-and-replace, so prose that happens to say "starter" — a README sentence, a comment — is left as written.
It gets its own secrets. .env is created from .env.example with a fresh
key for SECRET_KEY, JWT_SECRET and APP_KEY. A secret committed to a
starter is a placeholder by definition; no two projects should ever share a
signing key. An existing .env is never touched — it may hold real credentials.
It becomes a git repository, unless you pass --no-git.
Migrations, creating users, running the queue worker, starting the server — none of that is here.
Those belong to the project, in its own console.py:
python console.py db migrate
python console.py user admin ada@example.com ada
python console.py worker
python console.py serve --reloadThe framework provides the operations as plain functions —
sillo.record.commands, sillo.users.commands, sillo.work.commands — and
the project decides how to expose them. Nothing in a created project depends on
sillo-start, so it cannot rot when this tool changes, and you can delete this
tool the moment your project exists.
An archive member whose path escapes the destination is refused rather than sanitised — that is how a malicious archive overwrites files elsewhere on your machine.
A directory that is not empty is refused unless you pass --force, and even
then existing files are left in place.
A project name that is not a usable Python identifier is refused before anything is fetched, since it becomes both a package name and a directory.
See docs/contributing.md.
BSD-3-Clause.