This week, start a new project with Praat 7.0.02 if its workflow has no legacy dependencies; for an active study, keep the working version and test the upgrade in an isolated Mac environment before switching. This advice applies especially when scripts write files, call system commands, or depend on custom plugins and fixed configuration paths.
For students starting a speech study, this guide helps you decide whether to use the current Mac release as your baseline.
For researchers maintaining scripts, plugins, or automated jobs, it identifies the changes to test before migration.
For lab IT staff, it lays out a trial, rollback, and release process for a shared research environment.
Last updated October 9, 2026. Version and change details checked against the Praat change log and official Mac download page.
SECTION 01 What should you decide before upgrading Praat 7.0.02 on Mac?
Treat this as a workflow migration, not a routine application replacement. The official Mac download page lists Praat 7.0.02 and describes Mac availability for Intel and Apple Silicon systems, along with the system versions tested by the project. That confirms the release and its stated Mac scope; it does not certify your lab’s scripts, plugins, or research outputs. Check the official download information against the Mac you intend to use.
Your decision depends on what your project actually uses:
- New project with no inherited automation: trial Praat 7.0.02, then record the version and settings you use as part of the project baseline.
- Active project with scripts or automated file operations: preserve the working version and run a controlled regression test before changing the production workflow.
- Project with custom plugins, buttons, or saved preferences: inventory those items and their file locations before upgrading; confirm both migration and behavior.
- Cross-platform team: test on every operating system used to create, run, or review project files. A successful Mac launch is not a team compatibility result.
Praat 7.0’s official change notes describe script trust checks and changes to settings, button, and plugin locations. Those changes are reasons to test dependent workflows, not evidence that all older projects will fail. Keep that distinction in your lab notes: an official application change is confirmed; its effect on your project is something your team must verify.
Decision rule: If a project’s accepted outputs depend on an existing script, plugin, or saved configuration, do not replace its only working environment until a representative run passes and you can restore the prior state.
SECTION 02 Script trust checks require targeted retesting
The trust check matters most when a script does more than process data inside Praat. A workflow that creates or overwrites files, accesses paths outside its project folder, or invokes a system command has consequences beyond opening a sound file. Praat’s official change record documents the trust-check change in version 7.0; review the official change notes when deciding which scripts need a closer review.
Do not treat a confirmation prompt as a hurdle to bypass. First establish what the script is intended to do and whether its source is trusted. Then test it with a project copy and a public or de-identified sample. Your test should reveal whether the script can complete its intended work with the expected permissions, and whether it creates, modifies, or exports only the files you expect.
When a script calls Praat from a terminal or another automation layer, include the calling method in the test. The official manual documents calling Praat scripts from the command line; use that reference to check how your existing launch command is constructed. Review the actual command, arguments, working directory, input paths, and output paths on the Mac you will use. A script that runs from Praat’s interface may still behave differently when launched through a shell.
A useful script regression record includes:
- The script file and its source or revision identifier.
- The Praat version and Mac environment used for the test.
- The input audio and annotation files, with sensitive material removed where needed.
- The command or interface action used to start the script.
- The expected output files and the observed output files.
- Any permission decision, warning, or manual step required to complete the run.
If the script writes into a shared folder, test against a temporary project copy first. If it overwrites intermediate files, use a disposable output directory. Compare the resulting TextGrid, exported measurements, and logs with a known accepted run. Do not approve a migration solely because the script starts or because it produces a file with the expected name.
SECTION 03 Plugin, button, and settings migration needs separate validation
An application upgrade and a configuration migration are separate tasks. Finding a plugin in a new location proves only that you located or copied it; it does not show that the plugin works, that its dependencies are present, or that a button still calls the intended script.
Before moving an existing workspace, inventory what the lab has customized:
- Saved Praat preferences and any project-specific settings.
- Custom buttons and the scripts they call.
- Plugins, their versions, and any related external files.
- Scripts that refer to absolute paths, shared folders, or machine-specific locations.
- Written instructions that tell students where to install or select those items.
Praat’s manual has separate references for the preferences folder, plug-ins, and the buttons file. Use those references to identify the locations applicable to the version you are testing. Avoid copying an old folder wholesale into a new environment before you know which files are necessary: that can bring along stale settings and make it harder to identify what changed.
For each custom item, record three things: where it was stored before, where the new version expects it, and what user action or script uses it. Then migrate one item at a time in a test environment. Open the relevant menu or button, run its associated function on a copy of a known input, and check the result. If an item fails, you can then distinguish a path problem from a script, dependency, or plugin-compatibility issue.
Configuration note: A clean user profile can help distinguish a migration problem from an inherited preference. Keep the original profile untouched until the new profile has passed the project’s checks.
The plugin author’s adaptation is another separate dependency. Praat can provide a location for plugins, but it cannot guarantee that a lab’s custom plugin has been updated for a new release. If no maintained copy or test case exists, keep the old workflow available and ask the plugin maintainer or lab owner to review the code before relying on it for production analysis.
SECTION 04 How should you freeze and compare an active research workflow?
For an in-progress paper, thesis, or shared analysis pipeline, begin with a recoverable baseline. Preserve the version that currently produces accepted results, retain its relevant settings and custom files, and note how users launch the analysis. Do not make the new installation the only copy until your team has compared representative tasks.
Use a set of inputs that exercises the workflow rather than a single file that proves only that Praat opens. Include a representative audio file, its associated TextGrid, the scripts and plugins used in the study, and a known expected result. Where research data cannot leave a protected system, use a de-identified sample or a synthetic test case that still covers the same file and script operations.
Run the following comparison:
- Object handling: open the expected sound and annotation objects; confirm that the relevant objects and annotations are available.
- Annotation editing: make a controlled edit in a copy, save it, close the project, and reopen it to verify that the edit persisted as intended.
- Script execution: run the project’s actual analysis scripts and any relevant custom buttons, including command-line invocation where the project uses it.
- File operations: check that outputs appear in the expected locations and that existing inputs or intermediate files remain unchanged unless the workflow is designed to replace them.
- Exports and results: compare exported measurements, labels, and other expected outputs with the accepted baseline.
- Human review: have a researcher familiar with the analysis inspect the resulting annotation and output, especially where a file is created successfully but its content could still be wrong.
Keep a short result log beside the test inputs. Record the action, expected result, observed result, unresolved differences, and the person responsible for approval. A failed or unclear check is a reason to pause release, not to assume that a change is harmless. If results differ, identify whether the difference comes from the script, a changed setting, a plugin, a path, or a deliberate change in the research method.
This gives you three defensible outcomes. Upgrade when the required tasks pass and the lab has documented the new baseline. Defer when a required workflow fails and no validated workaround exists. Run both versions in parallel when new work can use the new release but an active study still depends on the previous environment.
SECTION 05 How can a cross-platform group approve the change?
A mixed Windows, Linux, and Mac team needs to test the handoff, not just the Mac installation. Check whether scripts assume a particular path separator, directory structure, shell command, or executable location. Identify which team member creates each project file, who runs the analysis, and which platform is used to review or archive the outputs.
Agree on one representative workflow, then run it on the platforms the team actually uses. Compare the delivered audio, TextGrid, scripts, and exported results at the handoff points in your process. If a script uses an external command, test that command on each relevant operating system. If the group shares plugins or scripts, confirm that each user can access the intended version and that local configuration does not silently substitute another copy.
The test should answer practical questions: Can a Mac-created project be opened and reviewed on the other team platforms? Do the scripts locate their inputs without manual path edits? Are outputs written to a location the next team member can access? Does a clean lab account behave like the account used during setup? Where the answer depends on a specific platform, document that limitation instead of describing the workflow as universally compatible.
For a new study, write the approved Praat version, script and plugin sources, relevant settings, and file conventions into the project’s environment notes. For an existing study, retain the old baseline until the project owner accepts the comparison. That is especially important when published or submitted results depend on a fixed analysis procedure.
SECTION 06 Mac acceptance timeline and release checklist
Use milestones that match your study’s risk. For a new project, a short trial before the first production analysis may be enough. For a long-running study or a shared lab installation, schedule acceptance before changing the environment used by the team.
- [ ] Baseline: record the current Praat version, settings, plugins, buttons, scripts, and launch method.
- [ ] Backup: preserve the working application and relevant configuration so you can return to the previous state.
- [ ] Scope: identify scripts that write files, call system commands, use fixed paths, or depend on custom plugins.
- [ ] Test data: choose a representative audio file, TextGrid, script, and expected result; use a safe copy.
- [ ] Isolated trial: install or open Praat 7.0.02 in a test environment without replacing the only working project setup.
- [ ] Migration review: locate preferences, plugins, and buttons using the version-specific official manual pages; migrate only what the test needs.
- [ ] Regression run: check object reading, annotation editing, file output, script execution, and human-reviewed results.
- [ ] Cross-platform handoff: repeat the required file and script checks on each platform used by the group.
- [ ] Release decision: approve the upgrade, retain the old version, or document a dual-track arrangement with an owner and rollback method.
For a new project, the release milestone is a documented, repeatable baseline. For active research, it is a passed comparison approved by the person responsible for the analysis. For a lab environment, it also includes instructions that let another researcher repeat the test without relying on undocumented setup steps.
If your group does not have a Mac available for this acceptance run, first review VPSNIX’s remote Mac environment options and the service help information to confirm that the access method fits your test workflow. A remote Mac can give you a separate environment for checking real macOS behavior without replacing the lab’s Linux or Windows systems; verify the needed access and workflow before you commit to a migration plan. You can also review available plans, but do not assume a particular configuration or cost without checking the current listing.
A remote Mac is not automatically the right long-term home for every study. If your team needs a permanently available local machine, physical peripherals, or sustained work that is better supported by existing lab hardware, use that hardware where it fits. But if the current alternative is borrowing a Mac with unpredictable access, delaying tests until a colleague is free, or making a costly hardware purchase just to validate one migration, a time-limited VPSNIX Mac environment can make the acceptance step more manageable. For Praat 7.0.02, the key is not where you run the test; it is whether you preserve the old baseline and verify the work your research actually depends on.