[DOCS] Name the on-device model in the domain doc - #113
Merged
Conversation
Prompt shaping arrived with rules but no vocabulary, and it puts a language model in an app whose whole story is that nothing leaves the Mac. The doc had neither the term nor the reason that story still holds. Two on-device models now exist here and they are not the same thing. A Speech Model turns speech into text and is the only one a default session touches. The On-device model rewrites text that already exists, through Apple's FoundationModels, and only with an explicit opt-in. Both belong to macOS, which is why the app is still a few megabytes. Three corrections came with it. The intro claimed the first milestone held only Direct Dictation with everything else deferred, which four shipped features ago stopped being true. "Version 1 inserts raw finalized text without filler-word or AI cleanup" sat two lines above the feature that rewrites it. And the no-bundled-models rule spoke only of speech.
jhampton
added a commit
to Swagatar-LLC/Talkify
that referenced
this pull request
Sep 3, 2026
The shaping work those notes described has now landed upstream (tornikegomareli#68, tornikegomareli#111, tornikegomareli#113, tornikegomareli#114) along with the insertion-clock fix (tornikegomareli#115), so keeping it in fork-unreleased.md would republish upstream release notes as if they were fork-only changes. The file accumulates again as fork work lands; empty is the correct resting state, and swagatar-release.sh already falls back to the build disclaimer when it is. Co-Authored-By: Vorno <agents-noreply@swagatar.co>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Prompt shaping landed with rules in CONTEXT.md but no vocabulary, and it puts a language model inside an app whose whole story is that nothing leaves the Mac. The doc had neither the term for it nor the reason that story still holds.
Adds three terms: Prompt Shaping, Shaping Prompt, and On-device model. The last one exists to keep a distinction the doc did not have: a Speech Model turns speech into text and is the only one a default session touches, while the On-device model rewrites text that already exists and only ever with an opt-in. Both belong to macOS, which is why the app is still a few megabytes.
Names
FoundationModelsin Architecture, with the reason the one-dependency rule and the privacy promise survive a language model: system framework, system model, no key, no account, no request.TranslationandNaturalLanguageare named alongside it as the other two frameworks that read the user's words.Three corrections came with it. The intro claimed the first milestone held only Direct Dictation with everything else deferred, which stopped being true four shipped features ago. "Version 1 inserts raw finalized text without filler-word or AI cleanup" sat two lines above the feature that rewrites it. And the no-bundled-models rule spoke only of speech models.
Docs only, no code. Full suite green.