OpenWand v0.11

Changed:

• Completed the OpenWand rebrand across the application, launchers, packages, documentation, website, translations, and release assets.
• Added automatic migration for existing Wisp settings, profiles, environment variables, and application data.
• Added preview-first, app-aware actions with explicit confirmation, target revalidation, verification, and recoverable execution.
• Added verified actions for PowerPoint desktop, LibreOffice Calc, and saved VS Code files.
• Added a full chat workspace with improved reply presentation, local formatting, code and LaTeX rendering, conversation management, and local forward sync.
• Added annotation-based rewrite review so proposed changes can be inspected before being pasted back into the active application.
• Added background Agent Team workflows with visible progress, task history, approvals, and reviewable results.
• Added shared and virtual workspace tooling with file activity, previews, recoverable file operations, and an Explorer-style interface.
• Added named model connections and clearer model routing across cloud providers, local models, Codex, and Claude.
• Added Cloudflare speech-to-text configuration, connection testing, and improved speech-provider status handling.
• Added prompt-injection warnings, clearer privacy summaries, and an optional first-prompt-only context default.
• Added new product demos and refreshed the English, Spanish, French, Simplified Chinese, and Traditional Chinese documentation.
• Added the experimental Formatted Replies addon for locally presenting model responses without an additional model call.

Fixed:

• Fixed packaged Linux startup by including the required EGL runtime dependency.
• Fixed false Windows packaged-shutdown failures by tracking managed process identity correctly.
• Moved packaged runtime and crash logs to the proper application data directory.
• Improved browser and application context capture, OAuth startup, speech initialization, and macOS safety behavior.
• Fixed chat scrollbar dragging, wheel scrolling, middle-button autoscroll, and reply-content preservation.
• Fixed repeated Settings menu creation and conflicting workspace activity updates.
• Stabilized Windows uninstall acceptance and packaged launcher testing.
• Expanded and stabilized cross-platform CI, release packaging, workflow coverage, and runtime acceptance tests.

Full Changelog: v0.10.2...v0.11

v0.10.2

Changed:

• Added full Claude model IDs and useful aliases to the OAuth model picker.
• Marked Claude Fable 5 as requiring usage credits.
• Updated Wisp package, runtime, client, and release metadata to 0.10.2.

Fixed:

• Exact Claude selections now persist and are forwarded unchanged to the Claude Agent SDK.

v0.10.1

Changed:
Added a combined local speech installer for Kokoro TTS and Whisper STT during onboarding.
Auto mode now installs GPU-capable speech packages while retaining CPU fallback.
Speech providers and settings are activated automatically after successful installation.
Added a “Restart app now” action when staged packages are ready to apply.
Updated translated installer messaging across all supported languages.
Fixed:
Fixed Auto mode incorrectly depending on unreliable GPU detection during installation.
Fixed speech settings being applied before package installation and verification completed.
Fixed staged speech installations not clearly prompting users to restart.
Fixed Kokoro Auto verification incorrectly requiring an NVIDIA GPU.
Fixed single-line custom prompts clipping glyphs and descenders.
Fixed translation tests inheriting personal caller settings.

v0.10.0

Changed:
Added native ChatGPT/Codex and Claude conversation harnesses with authentication, controls, persistence, documentation, and localized UI.
Added external conversation syncing and live agent conversation support across the overlay, chat, and runtime workers.
Added explicit local real-provider integration tests and expanded harness, transcript, intent-overlay, and conversation-sync coverage.

Fixed:
Isolated offline tests from personal provider and privacy settings.
Fixed macOS-incompatible Qt colors and completed translated harness guidance.
Prevented hosted CI from requiring provider API secrets.
Fixed expanding prompt sizing and deterministic text wrapping on Linux offscreen Qt.
Fixed Azure Artifact Signing setup when PSGallery is absent on Windows runners.

v0.9.3

Changed:
Added unified built-in and advanced local privacy filtering with a single translated review popup.
Added local privacy model installation and advanced privacy controls.
Added conversation file browsing.
Applied repository-wide Ruff cleanup.

Fixed:
Suppressed duplicate built-in findings when advanced privacy already blocks the same text.
Preserved intentional compatibility re-exports after lint cleanup.
Ignored the local pnpm store.

v0.9.2

Changed:
Added a first-run setup wizard with personal profiles, language selection, provider configuration, ChatGPT sign-in, and voice preferences.
Added secondary bindings and enable/disable controls for voice, dictation, context, screenshot, and other global hotkeys.
Added native Wayland screenshot capture and richer Linux accessibility context through AT-SPI.
Improved active-window, browser, selected-text, and document context handling.
Redesigned and expanded Settings, onboarding, overlay, bubble, and assistant workflows.
Improved local STT and Kokoro installation with pinned dependencies, progress reporting, install contracts, and Windows CUDA/cuBLAS support.
Added released-app speech diagnostics for troubleshooting local voice installations.
Added a complete self-uninstall workflow with translated UI.
Expanded Spanish, French, Simplified Chinese, and Traditional Chinese translations.
Added Azure signing for Windows release executables.
Expanded MCP documentation and clarified client/server responsibilities.
Fixed:
Fixed KDE screenshot delays, context-menu positioning, selected-text capture, and browser/application context detection.
Fixed stale or incomplete speech installations being treated as valid.
Fixed missing CUDA and cuBLAS components being detected too late.
Fixed CUDA warm-up and float16 fallback failures being silently suppressed.
Prevented STT and TTS installations from unexpectedly replacing shared dependency versions.
Fixed stale optional-package swap directories and unnecessary bundled pip files inflating releases.
Improved speech warm-up, shutdown, optional-dependency, and runtime stability.
Fixed secondary push-to-talk keys not receiving complete release-path test coverage.
Fixed release-manifest argument handling.
Fixed configuration tests leaking profile state from a developer’s local .env.
Fixed Windows-specific dependency tests failing on Linux CI.

v0.9.0

Added Wisp Context Server: trusted MCP clients can read selected text, clipboard, active window, browser-page text, and screen snippets.
Added ready-to-paste MCP client configuration and server setup documentation.
Fixed:
Prevented selection capture from accidentally copying text out of the assistant’s own window.
Serialized clipboard-based captures across Wisp and MCP processes to avoid clipboard restore races.
Added clear per-tool errors and startup diagnostics for missing capture dependencies.

v0.8.0

Changed:
Added Live Voice conversation mode with audio worker support, provider configuration, and settings controls.
Added a Live Voice hotkey path, including toggle/press-to-talk handling.
Added interleaved conversation captions in the overlay bubble.
Added optional google-genai install support for Live Voice.
Improved optional dependency install flows with staged installs, restart/resume handling, progress/status UI, and safer package probing.
Improved Linux browser/context capture using AT-SPI2, X11 owner detection, removable context rows, and stale-selection chips.
Updated localized UI strings for Live Voice, installer states, and related settings.
Fixed:
Fixed optional TTS/STT installs deleting unrelated packages in the shared google/ namespace.
Fixed CUDA Torch installs being overwritten during Kokoro setup.
Fixed Torch probe timeouts and fallback status handling during Kokoro install checks.
Fixed Kokoro TTS warmup/synthesis trying to go online when it should stay offline.
Fixed Live Voice readiness hints and push-to-talk modifier handling.
Fixed staged optional installs not preserving the Settings-captured wait PID.
Fixed stale tests after installer, overlay, and settings flow changes.
Fixed Linux browser context focus/tab URL detection issues.
Fixed stale X11 PRIMARY selections being reused incorrectly.
Fixed UI host shutdown lock fatal errors.

v0.7.2

Changed:
GitHub release builds now stay aligned with the same local portable build scripts.
Bubble notices can expand for long messages, start from the top, compact blank paragraph gaps, and stay visible longer based on message length.
Fixed:
Portable build drift between local builds and GitHub release artifacts is now guarded by workflow tests.
Long warning/status bubble messages no longer open scrolled to the end with earlier sentences hidden.
Tagged warnings, like hotkey or health failures, stay visible until dismissed instead of disappearing too quickly.

v0.7.1

Changed:
STT install now uses the same optional speech installer flow as TTS, with clearer progress, install status, model verification, and translated UI text.
Windows portable builds now bundle uv.exe for runtime optional package installs, including STT, Kokoro, and ElevenLabs.
STT and Kokoro installers now remove stale optional-package artifacts before repair installs.
Fixed:
Fixed STT installs appearing frozen or impossible to verify from Settings.
Fixed Setup Check reporting STT as healthy when faster-whisper existed but failed to import.
Fixed Windows portable builds failing when uv was missing or not discoverable after installation.
Fixed packaged runtime optional installs failing because uv.exe was not bundled.
Fixed broken STT/Kokoro rebuilds being reused instead of cleaned before reinstall.