Replace the HTML-drawn phone and keyboard mockups with Apple's iPhone 17 Pro
bezel and simulator recordings of the real keyboard extension (zh/en, light/dark).
Halve the marquee type and drop the credits image from the account section.
- Rebuild the story phone keyboard from AIKeyboardView: logo, AI/中文/EN
tabs, translation and clipboard chips, mic capsule states (dictation red,
hold-for-AI teal, busy spinner, inserted check), delete/space circles,
return capsule, and undo/edit after insertion
- Replace invented tone chips with the real three reply cards (ordinary,
formal, relaxed & playful) using the in-app demo copy
- Show the real copy flow: Paste plus ranked skill capsules in the top bar,
"Finding events…" and "Adding to Calendar…" status text
- Replace the outdated clipboard-skills video (retired circular chips) with
a live replica of the current top-bar skill flow
- Rework the hero capsule around the real mic capsule states
- Align FAQ, JSON-LD, and llms.txt wording with the shipped UI
- Move all canonical, hreflang, Open Graph, sitemap, robots, and llms.txt
URLs to the custom domain https://osg.osglab.com/
- Put category keywords (AI 语音输入法, 嘴替输入法, Agentic keyboard) in
the title, description, H1, section eyebrows, FAQ, and JSON-LD
- Add a 1200x630 social card and large-image Twitter card
- Add a scroll-driven three-act story (dictation, reply in your voice,
agentic action), a live voice capsule, split-text headings, number
counters, parallax media reveals, pointer spotlight and tilt, and a
circular theme transition
- Keep the page fully readable without JavaScript and under reduced motion
- Point legacy redirect pages at sections that exist
Lead with “Say it. It's typed. Speaks like you. Acts for you.” across the
README and product page: the personalization story (learned style,
three-tone replies) moves ahead of typing, clipboard skills are framed as
agentic actions, and typing becomes the fallback capability.
Replace the approximated flat render with the actual Icon Composer
export (transparent corners), and version the favicon URL since browsers
cache site icons aggressively.
- Render the new Icon Composer app icon for favicon, nav, and og:image
- Hero uses the three-device visual on a dark stage; hand-drawn mockups
replaced with on-device videos and real UI cards
- Add 2.1 selling points: personal style, three-tone replies, monthly
usage calendar, EventKit skills export, guided trials, credit rewards
- Restructure service paths with on-device as the primary trust card
- Self-hosted inline SVG icons; remove dead Three.js assets and logo.svg
- Bump page version references to 2.1.0 (Mac source at 2.2)
The speak-as-me skill inherited the Reply contract's JSON variants and
occasionally echoed the personal-style XML tags into the output. The
prompt now demands one ready-to-send plain-text paragraph, forbids JSON,
code fences, Markdown, lists, preambles, and multiple variants, and marks
the style block as prompt-internal packaging. Length follows the user's
natural voice instead of the mechanical ±20% cap.
Remove the test-mode unlimited personal-style corpus bypass and replace it with a real 1,250-character minimum gate. Production stays at 2,500. Training-corpus export and the diagnostics view now use a 5,000-character upper bound so the user's full private history never leaves the device in a single export.
- Add PolishStyleLearningCorpusBuilder.testBuildEffectiveCharacterCount (1,250) and trainingExtractionMaximumCharacterCount (5,000); trainingWindow(from:maximumCharacterCount:) accepts a custom cap.
- PolishStylesView: drop bypassesStyleLearningCharacterGate in favor of styleLearningMinimumCharacterCount, unify the ready/remaining text, remove the obsolete polishStyles.learn.testBuildReady key.
- PolishStyleCorpusExportStore: pass the 5,000 cap to trainingWindow and expose trainingExtractionMaximumCharacterCount in the export schema.
- FlowDiagnosticsSettingsView: show the export-sized (5,000-cap) corpus so the diagnostics match what the export ZIP actually contains.
- Tests: rewrite the export cap test for 5,000, add below-cap coverage, add trainingWindow(maximumCharacterCount:) unit coverage, and convert the old UI test into a regression test for the 1,250 test-build minimum.
- L10n: drop the testBuildReady string in en + zh-Hans, update diagnostics.corpus.ready to mention 5,000.
- CHANGELOG: bilingual entry under [Unreleased].
Removes assets and docs that are zero-references in the codebase and the
GitHub Pages site (scanned with ripgrep across *.swift / *.md / *.html).
Removed assets (no Swift / NSImage / UIImage / Image callers):
- 19 provider logo imagesets in OSGKeyboard/Assets.xcassets/:
anthropic, apple, ark, codingplanx, cometapi, custom, deepseek,
gemini, groq, mimo, minimax, mistral, moonshot, openai, openrouter,
qwen, siliconflow, xai, zhipu
- docs/assets/badges/appstore-{en,zh}.svg (README badges in use are
ios-{en,zh}.svg and macos-{en,zh}.svg)
Removed dead code:
- OSGKeyboardShared/Models/ProviderLogo.swift (no callers; only maps
provider ids to the now-deleted logo imagesets)
Removed placeholder screenshots:
- docs/screenshots/6.1/ and docs/screenshots/6.7/ (10 PNGs generated
by scripts/generate_screenshot_placeholders.py; never replaced with
real captures, no README/HTML/Swift references)
Removed zero-reference planning / audit md (no code, no README, no
CHANGELOG, no GitHub Pages HTML references):
- docs/ACCOUNT_MANAGED_GATEWAY_TESTING.md
- docs/APPSTORE_METADATA.md
- docs/STOREKIT_TESTING.md
- docs/TYPING_KEYBOARD_BASELINE.md
- docs/ios-pip-voice-session-plan.md
- docs/keyboard-accuracy-plan.md
- docs/keyboard-memory-budget.md
- docs/local-asr-architecture.md
- docs/mac-parity-audit-2026-08-23.md
- docs/personal-dictionary-icloud-kvs-checklist.md
- docs/screenshots/README.md (described the deleted placeholders)
Kept (still referenced):
- Brand marks: OSGBrandMark, OSGLogoWide, OSGStatusMark, osglogo
- docs/TESTING.md (referenced by README.md / README.en.md / CONTRIBUTING.md)
- docs/FLOW_BLUETOOTH_TESTING.md (referenced by docs/TESTING.md)
- docs/RELEASE_NOTES_1.8.0.md (referenced by CHANGELOG.md)
- docs/assets/{badges,whats-new,logo,app-icon,hero-devices} (GitHub Pages)
- AppStoreScreenshots/, marketing/ (App Store /策划素材, not in repo refs)
Updated .gitignore so the deleted md cannot be re-added locally and
pushed back by accident. Asset placeholders, dead code, and unused
badges are also covered by patterns under the same rationale.
No version bump: this is a pure cleanup (chore:) with no user-visible
behavior change.
- Add Generated/, CorpusRegistry, v6 boundary/blind evaluation
corpora, and sealed blessing benchmark review queues to the
generated-artifact ignore list so they stay out of the source
tree after reruns.
- Document the smart reply center, clipboard semantic v6 specialization,
personal style generation fix, and durable Apple account session
work in both English and Simplified Chinese.
- Update clipboard semantics README and open-training sources to
reflect the v6 boundary, blessing, and consensus-adjudication
corpora that the new release-gate pipeline consumes.
- Add reproducible v6 boundary, blessing, and consensus-adjudication
corpora, plus the tiny-transformer trainer and v6 release-gate
evaluator that gate every candidate on the deployed baselines.
- Wire consensus-label merging, product-policy anchor evaluation, and
sealed blessing benchmark review with their pytest coverage.
- Refresh open-training corpus generation, iterative retraining runner,
and random-holdout evaluation so v6 candidates can be benchmarked
end-to-end.
- Cover reply-variant decision paths, scene-detector abstention, and
the merged reply center stance catalog.
- Lock in clipboard semantic analyzer thresholds, complaint-only
suppression, and existing-skill routing.
- Add durable session, refresh, and managed-gateway reliability tests
for the Apple account path.
- Verify personal-style prompt derivation, low-confidence ASR
tendencies, and Polish style learning against real corpus evidence.
- Exercise the assistant keyboard and Polish styles UI flows in
end-to-end UI tests.
- Merge invitation, task, blessing, clarification, and empathy actions
into a single Reply flow, with three fixed, clearly labeled stance
choices whenever user intent must not be guessed.
- Refine clipboard semantic routing with bilingual schedule,
confirmation, and follow-up models, conservative language thresholds,
and explicit-assignment guard for complaint-only text.
- Persist Apple account refresh state, harden session recovery, and
surface durable account diagnostics across keyboard and app.
- Derive personal-style prompts through two-stage corpus evidence and
apply real low-confidence ASR tendencies instead of neutral templates.
- Localize the new reply center, clipboard semantics, and personal-style
surfaces in both English and Simplified Chinese.
Resolve the pinned XcodeGen archive layout and let the keyboard derive bounded clipboard recommendations from the full catalog independently of installation state.
Unify PiP recovery across startup and foreground transitions, surface actionable status, and add privacy-safe analytics plus the refreshed onboarding and account experience. Update documentation assets and advance the release build to 84.
Introduce optional Apple account-backed credits with scoped gateway access while preserving local and BYOK paths. Refresh assistant behavior, tests, privacy disclosures, docs, and the website for the 2.0 experience.
Merge dictation and AI controls into one assistant surface, preserve safe insertion and clipboard actions, and align settings and tests with the new flow.
Replace the unclear CLM source, align release and privacy disclosures, complete third-party notices, and keep the expanded test suites deterministic for release validation.
Keep English QuickType/mmap and Chinese schema hygiene on one branch.
Secure fields skip Rime; English still loads its lexicon only while English is active.
Erase dialect syllables before abbrev so wom prefers 我们, flush librime
on leave, and add a settings action to forget learned Chinese and English order.
Add save-to-Notes export and direct map handoff while aligning multi-touch typing, period shortcuts, return actions, and navigation visuals with system behavior.
Rotate AI idle suggestions with optional remote packs, move history/dictionary onto self-sizing Home preview cards, harden clipboard capture/prompting, and simplify keyboard chrome by dropping most liquid-glass shadows.
Ship optional keyboard clipboard history with settings, suggestion strip, and paste-permission guidance; let undo roll back clipboard inserts; bump build to 64.
Add the AI keyboard tab, Agent settings, and user-owned LLM key path for 1.7.0, including streaming answers and web-search transports without the built-in DeepSeek fallback.
Adapt typing/voice surfaces for iPad width and height, add the system globe
key and last-input editing flow, harden host-only Rime deployment, and remove
clipboard voice commands. Bump build to 61.
Keep marketing version at 1.6.6 and bump build to 59. Strengthen never-answer
polish safeguards, clipboard reply-intent continuity, voice undo UX, device
UITests harness, eval fixtures, and What's New assets.
Extend the sheet WebView under the home indicator, keep scroll insets cleared after navigation, and neutralize remote page safe-area bottom padding in-app.
Stabilize clipboard long-press prepare/resume across paste alerts and cold start, add an in-app release-notes sheet with remote bilingual HTML, localize typing input settings, and bump build to 55.
Long-press mic runs ASR as an instruction over eligible clipboard text,
with host-confirm recording UI, min-record gate, and light ASR prewarm.
Bump CURRENT_PROJECT_VERSION to 53.
Custom polish styles can opt in to emotion-matched emoji (default off), with
prompt-level opt-in detection so paste-only styles keep model-added emoji.
Also include Volcengine API-Key ASR auth, voice-processing capture, PiP flash
fix, and related keyboard Shift/haptics reliability work.
Redeploy an osg_personal import table on dictionary add/delete/sync so
Chinese, English typing, and ASR share one curated lexicon (next keyboard open).
Rocky