Complete Wave 0 launch foundation and shell decomposition - #23
Open
Mvkd108 wants to merge 74 commits into
Open
Conversation
|
Important Review available on request
Reviews should be triggered manually for repositories with fewer than 10 stars. Select Trigger review above or comment ⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Adds ignore rules for .NET build/test/coverage output, downloaded model weights and native runtimes, local recordings and qualification results, signing material, packaged installers, and agent-local scratch files. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds the full license texts referenced by THIRD-PARTY-NOTICES.md so the distributed package can ship attribution offline. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Captures the phased delivery plan for the Windows native app and tracks feature parity against the macOS build. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…vice Introduces AtomicJsonFile for crash-safe JSON writes, CrossProcessFileLock for coordinating access across app instances, and MuesliPathService as the single source of truth for app data locations. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Stores API keys encrypted at rest with the current user's Windows DPAPI key rather than in plaintext settings. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Ensures only one Muesli instance owns the hotkeys and capture devices, handing activation to the running instance instead of starting a second. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Centralises monitor-aware window positioning and restores placement across DPI and display-topology changes. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds NativeSherpaRuntime to resolve and load the bundled sherpa-onnx native libraries (CPU and CUDA), SafeArchiveExtractor for path-traversal safe model unpacking, and ModelSetupArtifactCleaner to reclaim partial downloads. Wires the sherpa-onnx and SharpCompress package references and ships the CUDA runtime manifest. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Introduces a common TranscriptionProgress contract plus a throttling wrapper so long transcriptions can report progress without flooding the UI dispatcher. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds the sherpa-onnx Parakeet client for streaming recognition, a timestamp segmenter that turns token timings into readable segments, and a PCM normaliser that converts capture buffers to the model's expected sample rate and format. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Describes the supported ASR models, their download sources, on-disk layout and hardware requirements in one place so download, qualification and UI all agree on what is installable. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds the batch (non-streaming) recognition path used for file imports and post-meeting finalisation, sharing model resolution and progress reporting with the streaming client. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Fronts the streaming and offline clients behind a single entry point that selects the right backend from the catalog and the user's configured model. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Handles install, verification, repair and removal of ASR models, and resolves which streaming platform (CPU or CUDA) a given model can run on for the current machine. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds sherpa-onnx segmentation and embedding based diarization, plus a service for persisting user-assigned speaker names across a meeting. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds a LLamaSharp-backed cleanup pass that fixes punctuation and casing locally, a filler-word filter, and the pipeline that sequences raw ASR output through filtering and cleanup into final text. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Probes the installed native runtime, models and hardware at startup and maps the result to a user-facing status so the UI can explain exactly what is missing. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds sherpa-onnx segmentation and embedding based diarization, plus a service for persisting user-assigned speaker names across a meeting. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Measures accuracy and latency across installed models on a fixed corpus, and bounds how long benchmark captures are retained on disk. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…ecks Centralises where captured audio lives and when it is deleted, adds an explicit retention policy for dictation scratch audio, surfaces the Windows microphone privacy setting, and adds per-process loopback capture. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…chine Replaces ad-hoc hotkey handling with an explicit gesture type, a probe that detects conflicts with other applications, and a state machine that models push-to-talk and toggle dictation deterministically. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds parsing for Zoom, Teams and Meet URLs, resolution of the most likely active meeting from open windows and calendar entries, presence signals from process and audio state, and platform badges for the UI. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…nal store Models a meeting session as explicit states with legal transitions, keeps an ordered transcript timeline, records live-transcript window ownership, and journals session state to disk so a crash mid-meeting can be recovered. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…th monitoring Drives streaming recognition for the duration of a meeting, detects and backfills gaps where the stream dropped, and monitors capture health so silent or dead devices are reported rather than silently producing nothing. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Declares the audio and video containers accepted for transcript import and maps decoder plus recogniser progress onto a single progress bar. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…ider disclosure Composes structured notes from transcript and summary output, adds playback of the meeting recording aligned to the transcript, and states plainly which provider processed a summary and whether it left the device. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Runs an end-to-end readiness check over runtime, models and capture devices before a meeting starts so failures surface up front. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Runs the configured follow-up actions once a meeting finalises: export, summarise and hand off notes, with each step independently recoverable. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds the planner that turns a requested desktop action into a reviewable step list, a trace store recording what was actually executed, a UI bridge for surfacing and confirming steps, and an out-of-process automation test host so the flow can be exercised without driving the real desktop. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…ction and notes Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…ding Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Git checks source out with CRLF on Windows (core.autocrlf=true), so the literal "\n" in the Show()/PositionWindow ordering assertion never matched on a clean clone and the test failed for reasons unrelated to the invariant it guards. Normalise the file text before asserting. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…t merge reapplication
Wires the dashboard, dictation, meetings, model management, benchmarking, onboarding, computer-use and diagnostics surfaces onto the native services introduced in the preceding changes. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Drops the Python worker payload from the package, bundles the native runtime and license texts, and updates the Inno Setup script plus the install, uninstall, signing and package smoke-test scripts to match. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Adds scripts to smoke-test each catalogued model, install the Parakeet CUDA runtime, and benchmark native transcription and meeting throughput. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Runs dictation against a fixed corpus and per-target suites, and summarises latency percentiles for release sign-off. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…cripts Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Builds Release, runs the hardware-free regression suite, validates PowerShell script syntax, and builds and smoke-tests the public win-x64 package on every pull request. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Updates the READMEs, roadmap, native plan, benchmark results and v1 release notes to describe the in-process sherpa-onnx pipeline that replaced the Python worker, and records the context-transfer notes. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
…t merge reapplication
Wires the dashboard, dictation, meetings, model management, benchmarking, onboarding, computer-use and diagnostics surfaces onto the native services introduced in the preceding changes. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Custom ApplicationManifest was replacing the SDK-generated DPI block, so tests now extract RT_MANIFEST from the built EXE and assert PerMonitorV2 plus the true/pm fallback. Co-authored-by: Cursor <cursoragent@cursor.com>
Public notices claimed a CUDA provider that the zip does not ship. Align notices, licenses, native inventory, and diagnostics so packaging tests fail on CUDA disclosure/runtime disagreement. Co-authored-by: Cursor <cursoragent@cursor.com>
Give FeatureRuntime a repository-facing boundary without wiring AppServices or the locked runtime files. Co-authored-by: Cursor <cursoragent@cursor.com>
Lock load/save order, missing and corrupt files, ParentId vs one-level folders, and the absence of a fake migrated-success path. Co-authored-by: Cursor <cursoragent@cursor.com>
Map every production AppDataStore call, freeze the Wave 1 file lock, and specify rollback, schema, and ParentId rules without marking L27 complete. Co-authored-by: Cursor <cursoragent@cursor.com>
Keep MTG-03 Partial: save/cancel and accept/reject cannot destroy the prior transcript, while FeatureRuntime wiring remains a later integration slice. Co-authored-by: Cursor <cursoragent@cursor.com>
…ication The human corpus and four paste reviews cannot be fabricated. This adds a schema-2 placeholder template that fails ValidateOnly, rehearsal tooling that reports qualified=false, and rejection coverage for empty/model-only references. Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Mvkd108
force-pushed
the
codex/wave0-launch-foundation
branch
from
August 18, 2026 15:42
d5032cf to
3bff598
Compare
|
Claude encountered an error after 0s —— View job I'll analyze this and get back to you. |
|
Review the following changes in direct dependencies. Learn more about Socket for GitHub.
|
|
Warning Review the following alerts detected in dependencies. According to your organization's Security Policy, it is recommended to resolve "Warn" alerts. Learn more about Socket for GitHub.
|
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.
Summary
Verification
The four skips require absent real media, streaming-model, or multi-speaker fixtures and are reported explicitly rather than silently passing.
Need help on this PR? Tag
@codesmith-botwith what you need. Autofix is disabled.