Repository navigation
Snapmaker Studio v1.7.0 — One job screen, print risk signals and a colour preview
Independent open-source project — not affiliated with or endorsed by Snapmaker.
"Snapmaker" is a trademark of its respective owner.
These notes describe the build named in
RELEASE_METADATA.md,
published unchanged from release-candidate run 40129442382. The previous release is v1.6.0.
This release puts a project on one job screen, replaces the readiness percentage with a list of the risks Studio
found, draws a read-only colour preview, tightens the checks that decide whether a project fits the plate, and
updates the update check's TLS dependencies.
One job screen
- Opening a project lands on one screen: the model in the middle, the target, materials and copy settings in a
pane beside it, and a single button for the next step (create the U1 copy, open it in Snapmaker Orca, compare the
slice). The copy's file name and folder, and the settings comparison after slicing, appear in the same place. - Opening or restoring a job never prepares, opens Orca or slices by itself, and every existing check still applies.
A colour preview
For a millimetre 3MF, Studio draws the model by the filament that prints each part, links each filament to its
toolhead, and lets you compare the project's own colours, the spool colours you have chosen so far, and the colours
read back from the copy Studio wrote. It is read-only. More than four colours stay visible, each in its own colour;
Studio does not remove or merge them. Surface the project does not assign is shown grey. When a preview cannot be
built (another unit, an unreadable file, too much detail, no 3D graphics) Studio says why and shows the plate
instead.
Print risk signals replace the readiness percentage
The "Print readiness (estimate)" percentage, the star rating and the Studio score are gone. Studio now lists each
risk it found, once per condition, with what it means and what to do, and says what it did and did not check.
Community guidance attaches only to the conditions it actually fits.
Placement and size checks say what they did not check
- Bed fit reports each object's size and the instances' placement as separate facts; several small, separated
objects are no longer summed into one "too big" object. - A copy of a repeated object that hangs off the plate is no longer hidden by one that does not.
- Placement and plate-fit checks are reported as not checked for projects whose physical size the file does not
establish (a non-millimetre unit, or a rotated or scaled object). - A project with several plates is no longer called ready just because each plate fits on its own.
- Studio no longer recommends Arrange all plates as a fix; it says to review the project in Snapmaker Orca before
slicing. - When the bed-fit check could not run, the report says it was not verified instead of reading as a clean result.
- Each setting Studio adjusts when it prepares a copy says why, and labels whether that statement is Studio's own
check or advice to verify the setting in Snapmaker Orca. The labels make no claim about how Snapmaker Orca behaves. - The cost line in the Business Doctors says "estimated cost".
Security
- The update check's TLS dependencies are updated (rustls 0.23.46, rustls-webpki 0.103.16), and it no longer trusts
theemSign ECC Root CA - G3root that webpki-roots 1.0.10 removed (RUSTSEC-2026-0381, the October 2026 eMudhra
disclosure). - The update check only follows
releases/latestandreleases/tag/<tag>links of this repository. - If you build Studio from source, Node.js 22.23.4 or a newer fixed line is required.
This does not change what is already installed: people running an earlier release still have its binary until they
install a newer one.
What was checked, and what was not
The candidate was built once by the project's release-candidate workflow. Its Windows installer and Linux .deb
were re-downloaded and checked by SHA256 against the build manifest. The Windows installed-application acceptance
run used an acceptance copy of the installer (same application files, separate install identity) on a clean
local account and passed 54/54; that run includes an upgrade from an acceptance copy of v1.6.0, which is not a
production upgrade. The real installer's own install, upgrade and uninstall were exercised only on a disposable
GitHub-hosted runner. The Linux package passed its in-container acceptance run and clean-environment validation on
Ubuntu 22.04 and 24.04 (49/49 each).
Real printer: read-only verification against a real Snapmaker U1 passed 76/76 on its second attempt, using real
Spoolman and Bambuddy containers with synthetic spools. The first attempt failed on defects in the verification
tooling, not the product; the tooling was corrected and reviewed before the second. Snapmaker Orca: a painted
project and its prepared copies loaded in an isolated Orca 2.4.1 profile; the bundled sample project (examples/demo_u1_showcase.3mf) would not load in Orca 2.4.1 even
unprepared (known issue: the file is a minimal synthetic package whose content-type and relationship parts are empty stubs, which isolated testing confirmed is what Orca 2.4.1 rejects; the file shipped in this release is unchanged), so it proves nothing about its prepared copies. Not checked: any real SpoolEase device.
Studio is advisory. It does not slice, it does not promise a successful print, and anything sent to your printer
is something you confirm.
Verify your download
These are the values of the v1.7.0 downloads on this release.
- Windows:
Snapmaker.Studio_1.7.0_x64-setup.exe— 22,748,251 bytes — SHA25671c91b0c2cf2ec41d80bc75bff828ff6b35d6ff3fd93764b894676ec9d360b04 - Linux:
snapmaker-studio_1.7.0_amd64_210d86868882.deb— 26,325,960 bytes — SHA2562af5ceb2f6b08e4f732df8457924ea7e9135c7f9408dda41c688c189298ccabc
These values repeat the canonical record in
RELEASE_METADATA.md on purpose:
this file becomes the GitHub release body, and a downloader must be able to verify a hash without opening a second page.
Neither installer is code-signed yet; verify the hash before you run it. Verification status for this build is
in TRUST_STATUS.md.