Official Python client for the System Locker Simple protocol
(POST /auth): one request, one answer. No sessions, no heartbeats, no
signatures — the right fit when the machine running the check is one you
control. For software distributed to untrusted machines,
use a Bedrock client instead: it verifies an Ed25519 signature on every
response.
pip install systemlocker-simpleZero dependencies — standard library only. Python 3.10+. Fully typed.
from systemlocker_simple import Client, Config
client = Client(Config(
system_id="abcdefghijklmnopqrst", # from the dashboard
version="1.0.0",
# hwid stays "1" unless you want device locking.
))
try:
ok = client.authenticate_with_key("SL-XXXX-XXXX-XXXX")
except Exception:
raise SystemExit(1) # transport / server error — never counts as licensed
if not ok:
raise SystemExit(1) # rejected — block the action
# …run the gated action…| Operation | Method |
|---|---|
| Check a license key | authenticate_with_key(key) |
| Check username + password | authenticate_with_password(user, password) |
Key expiry (Never or a UTC date) |
key_expiration_for_key / key_expiration_for_password |
| Server-side variable | get_variable(name, key="") |
| Self-service HWID reset | reset_hwid_for_key / reset_hwid_for_password → ResetOutcome.GRANTED/DENIED/TOO_SOON |
Errors raise SimpleError whose kind separates infrastructure problems
(Transport, Server) from license denials (Denied with the server's raw
reason: frozen, hwid banned, expired key, …), Google-SSO cases (SSO,
with sso_link(error) for the portal URL), LOCAL_FAILURE when an opt-in
SL-HWID device identity cannot be produced (see below), and unrecognized
reasons (UnknownReason, raw string carried).
client = Client(Config(system_id=…, version="1.0.0", api_key="…"))
count = client.management().redeemed_user_count()
keys = client.management().generate_keys(EXPIRY_ONE_MONTH, 10, "june-batch")Wraps POST /api/v1: key status/expiration, HWID resets (single/admin or
whole-system), key generation, bans, expiry adjustment. Keep the API key on
servers you control.
Accounts created through Google sign-in have no local password on the
server. A username/password check for such an account fails with an
sso, ssoexp, or ssowrong reason that embeds the portal URL where the
user completes Google sign-in and receives a system-specific password
(valid 180 days) to use as their account password. There is no callback;
the user transcribes the generated password into your login form and you
simply retry.
Deliver the portal link to your user through your own channel (API response, email, chat).
from systemlocker_simple import ErrorKind, SimpleError, sso_link
try:
client.authenticate_with_password(username, password)
except SimpleError as error:
if error.kind is ErrorKind.SSO:
# sso / ssoexp / ssowrong — the portal URL is embedded in the error.
portal = sso_link(error)
send_to_user(user, portal) # your channel: API response, email, chat…
return
raise # any other denialgoogle_sso_url(system_id) (or client.google_sso_url()) builds the same
portal URL before any denial, if you already know the account signs in
through Google.
The default derivation is a plain hardware hash:
from systemlocker_simple.hwid import device_hwid
config.hwid = device_hwid()Derives a stable identifier from the machine GUID, hardware UUID, CPU id,
and MAC (Windows and Linux). It stays available, but it is the weaker
option: the hash over-fits a handful of hardware values, so swapping a disk
or NIC — or cloning the machine into a VM — changes the HWID and forces your
user through a device reset. A developer-supplied stable value works just as
well. Set config.hwid = "1" only to explicitly disable device locking.
config = Config(hwid_mode="sl-hwid") # default is "legacy"SL-HWID derives the HWID from a random key locked behind threshold secret sharing instead of hashing hardware directly. It is fault tolerant and cross platform (Windows, macOS, Linux), combines 14 hardware factors, and any two of them can fail or change without changing the HWID; drifted factors are quietly re-absorbed after each successful authentication. The module's own persisted value is hard-locked, so copied state cannot stand in for changed hardware.
Things to know before enabling it:
- The HWID changes. SL-HWID produces a new opaque identifier, so a
deployment switching from the legacy hash (or a custom value) must reset
its claimed HWIDs once, at rollout — per key, self-service, or
system-wide through the management API — or users will hit
hwidmismatches. - Storage is shared. The enrollment lives in one per-machine location
(the registry on Windows, an application-support directory elsewhere),
shared by every System Locker client on the device. Configure
sl_hwid_storeonly when you deliberately need separate device state. - Re-activation exists. If hardware drifts past the recovery threshold,
requests fail with a
LOCAL_FAILUREerror and the user needs a reset.
SL-HWID is the right choice for a launcher: a Simple-based launcher that
opens a Bedrock-protected program reports exactly the HWID the Bedrock
client reports, because both share the same per-machine enrollment. The key
the user already activated in the launcher works for the protected program
too — one device, one HWID, no hwid mismatch between the two.
SL-HWID changes the device identifier only. It does not change what the Simple protocol guarantees: responses are still unsigned, so only use this client on machines you control.
An explicit hwid value (including "1") always wins over both modes.
See SECURITY.md. Report vulnerabilities privately through the System Locker support channels, not via public issues.