AnswerCue v2.8.11

Release date: 2026-10-08

Summary

AnswerCue v2.8.11 saves preparation chats as interviews as soon as the first message is sent. Conversations appear immediately in the interview sidebar and receive automatic titles without requiring a live interview first.

What's New

• Chat interviews: Sending the first message creates one saved interview and updates the sidebar count immediately.
• Automatic chat titles: The selected LLM generates a short title from the first message. The sidebar and conversation header update together.

Improvements

• Prep conversations remain available after reopening the app or switching interviews.
• Starting a live interview from a saved prep chat reuses the same interview ID. Stopping and restarting continue the same history.
• Chat-only conversations show Start interview and remain in the preparation phase until a live session is started.

Fixes

• Prep chats no longer remain hidden from the interview list until a live session ends.
• Empty drafts do not create sidebar entries. Repeated messages and save retries do not create duplicate interviews.
• Delayed title generation cannot overwrite a manual rename or the title of a completed live interview.
• A title request failure retains the chat with a title derived from the first message.

Technical

• Uses the existing SQLite meeting index and workspace storage; no database schema or storage-path change is required.
• Saves the conversation before launching asynchronous title generation. Title generation uses the selected LLM provider and may incur a small additional API request.
• Full automated suite: 1,062 passed, zero failed, 41 skipped. Adds regression coverage for first-message registration, empty drafts, duplicate retries, title failures, and manual/live title protection.
• Browser UI checks cover sidebar/header title updates, subsequent messages, app reopening, and reuse of the saved interview ID on live start.
• Existing provider-key and installed-user audio checks still require user-machine testing. Electron backend type checking retains pre-existing diagnostics.
• Release CI validates packaged Mac runtime dependencies, DMGs, signatures, and macOS 27 Dock transitions. Windows CI verifies package inputs, the packaged renderer, and installer signing status.
• User configuration, provider keys, Persona, custom instructions, and previous interviews are retained.

Platform Downloads

macOS

• Apple Silicon: AnswerCue-2.8.11-arm64.dmg or AnswerCue-2.8.11-arm64-mac.zip.
• Intel: AnswerCue-2.8.11-x64.dmg or AnswerCue-2.8.11-mac.zip.
• Auto-update metadata: latest-mac.yml and ZIP blockmaps.

Ad-hoc signed and unnotarized. Replace the earlier app in Applications. If Gatekeeper blocks this trusted fork build:

xattr -dr com.apple.quarantine /Applications/AnswerCue.app

Windows

• Windows x64: AnswerCue-Setup-2.8.11.exe.
• Auto-update metadata: latest.yml and installer blockmap.

Unsigned when Azure signing credentials are absent. Windows may display an unknown-publisher or SmartScreen warning.

Linux

No Linux package is included in this release.

AnswerCue v2.8.10

Release date: 2026-10-08

Summary

AnswerCue v2.8.10 updates the desktop runtime and stabilizes macOS application/helper identities to reduce Dock registration churn. It addresses the reported disappearing duplicate icons while retaining the lean package, cloud transcription, and persistent interview features.

What's New

• Current Electron runtime: Updated from Electron 33.4.11 to 44.7.0 for current macOS behavior.
• Golden Gate release validation: CI tests the packaged app on macOS 27, including a real BrowserWindow and repeated native Dock hide/show transitions without recording audio or calling provider APIs.

Improvements

• Keeps macOS app identity separate from disguise presentation. Explicit disguises still change process/window presentation and available icons; they no longer rename the Mac application or inject a CFBundleName environment override into child processes.
• Retains electron-builder's canonical helper names, bundle identifiers, and background flags instead of changing helper bundle names to CoreServices.
• Initial normal mode uses the packaged Dock icon. Repeated identical disguise requests are idempotent, and normal mode does not schedule repeated process-title refreshes.
• Keeps the architecture-specific dependency allowlist and packaged runtime checks introduced in v2.8.8.

Fixes

• Updates native rebuild/ABI tooling and the SQLite addon for Electron 44’s V8 API. Removes the deprecated duplicate electron-rebuild package. Database storage paths and schema are retained.
• Removed repeated Mac application-name and default-icon updates during startup.
• Removes the inherited CFBundleName override before creating the normal app windows and helpers.
• Adds regression checks for stable Mac identity and idempotent disguise/icon behavior.

Technical

• Bumped package version to 2.8.10; source and release notes are public in this fork.
• The user's diagnostic showed one foreground AnswerCue process and one accessory plugin helper. The extra icons disappeared when clicked. This rules out three registered foreground instances and points to stale Dock registrations, but does not prove which operation originally created them.
• The original v2.8.8 launch/reopen tests showed one Dock entry on clean macOS 14, 26, and 27 runners. The debugger-based visibility test was inconclusive; this release adds an isolated packaged native visibility harness.
• Full automated suite: 1,057 passed, zero failed, 41 skipped. Renderer/production builds, packaged database/keytar/audio loading, image conversion, PDF/DOCX/TXT imports, per-architecture inventory, signatures, DMGs, and updater ZIP checks remain release requirements.
• Electron type checking retains existing fork diagnostics. Live provider-key transcription and installed-user microphone/system audio still require user-machine checks.
• Clean-runner validation cannot guarantee repair of a Dock cache already corrupted on another Mac. Existing ghost entries may need a one-time killall Dock after replacing and quitting the old app. This update avoids the unnecessary identity mutations that can contribute to future registration inconsistencies.
• User settings, keys, custom instructions, Persona, and interview history use the existing storage paths. No global Dock preferences or other applications' history are changed by the app.

Platform Downloads

macOS

• Apple Silicon: AnswerCue-2.8.10-arm64.dmg or AnswerCue-2.8.10-arm64-mac.zip.
• Intel: AnswerCue-2.8.10-x64.dmg or AnswerCue-2.8.10-mac.zip.
• Auto-update metadata: latest-mac.yml and ZIP blockmaps.

Ad-hoc signed and unnotarized. Replace the earlier app in Applications. If Gatekeeper blocks this trusted fork build:

xattr -dr com.apple.quarantine /Applications/AnswerCue.app

Windows

• Windows x64: AnswerCue-Setup-2.8.10.exe.
• Auto-update metadata: latest.yml and installer blockmap.

Unsigned build. Windows may display an unknown-publisher or SmartScreen warning. Azure signing remains available when all signing credentials are configured.

Linux

No Linux package is included in this release.

AnswerCue v2.8.8

Release date: 2026-10-08

Summary

AnswerCue v2.8.8 reduces Mac package size by removing unused libraries, shipping compiled frontend/backend code without duplicate SDK copies, and selecting native libraries for the target architecture. It retains the cloud transcription and interview features from v2.8.6.

What's New

• Packaged runtime verification: Mac CI starts the built app in an isolated diagnostic mode and checks the native database, key storage module, audio addon, image conversion, PDF/DOCX/TXT imports, and transcription module loading without recording audio or calling provider APIs.
• Package inventory checks: CI rejects missing runtime files, development libraries, source maps, and native modules belonging to the other Mac architecture, and reports the largest packaged components.

Improvements

• Frontend libraries and bundled JavaScript SDKs remain available for development and compilation but no longer ship as duplicate node_modules packages.
• Keeps only native/asset-backed runtime libraries and dynamic validation/authentication helpers in production dependencies.
• Removed unused ElevenLabs browser client, Google Stitch, Tavily, sqlite3, liquid-glass-react, three, tap, screenshot-desktop, unused styling/code-block utilities, and obsolete type packages. Legacy OCR remains development-only for its existing tests.
• Production Electron code is minified. Developer builds keep their existing layout and source maps for tests and debugging.
• Native audio, Sharp/libvips, and PDF canvas libraries are selected per Mac architecture. Both architectures are prepared during CI; each app ships only its matching files.
• Packaging includes runtime icons rather than the full source/readme artwork directory, and excludes unused PDF browser/ESM builds and dependency source maps.
• Compiled code retains generated third-party notices and the project license. The license for locally adapted glass-effect code is retained after removing its npm package.

Fixes

• Mac file rules now retain the production allowlist instead of replacing it with exclusion-only rules, which previously caused the candidate to include source/build files and both audio architectures. Tests exercise the complete electron-builder matcher setup; CI rejects source, temporary, build, and environment files.
• PDF imports prefer the unpacked worker file so Node can load it from a real filesystem path.
• Ensures both Mac canvas variants are installed during cross-architecture builds, alongside Sharp's native image libraries.
• Local release commands use the same production Electron bundling mode as CI.

Technical

• Bumped package version to 2.8.8. Source and detailed release notes are published in this fork.
• Full automated suite: 1,055 passed, zero failed, 41 skipped. Renderer type checking and production builds pass. Electron type checking retains the unchanged fork's 37 diagnostics, with no new diagnostics.
• File-selection tests exercise electron-builder's actual filters for both Mac architectures, including native library retention and rejection of other architectures.
• Mac release CI must pass packaged runtime probes, dependency inventories, signature checks, mounted DMG verification, and extracted updater ZIP checks before release assets are published. Intel runtime execution requires Rosetta on the build runner; CI explicitly reports a skip if unavailable.
• Electron's Chromium/Node runtime, native audio, database, key storage, image libraries, and PDF parsing remain required. No local AI model is included or downloaded.
• Existing provider keys, custom instructions, Persona, and interview history use the same storage paths. No user data or old model cache is deleted.
• Live transcription with user API keys, microphone/system audio capture, macOS permissions, and Dock behavior still require installed-Mac checks. Packaging probes do not record audio, read keys, or make paid API requests.

Platform Downloads

macOS

• Apple Silicon: AnswerCue-2.8.8-arm64.dmg or AnswerCue-2.8.8-arm64-mac.zip.
• Intel: AnswerCue-2.8.8-x64.dmg or AnswerCue-2.8.8-mac.zip.
• Auto-update metadata: latest-mac.yml and ZIP blockmaps.

Ad-hoc signed and unnotarized. Replace the earlier app in Applications. If Gatekeeper blocks this trusted fork build, remove its quarantine flag:

xattr -dr com.apple.quarantine /Applications/AnswerCue.app

Windows

No Windows installer is included in this release.

Linux

No Linux package is included in this release.

AnswerCue v2.8.6

Release date: 2026-10-08

Summary

AnswerCue v2.8.6 replaces local speech transcription with selectable OpenAI and ElevenLabs real-time transcription. LLM settings, transcription keys, and the active transcription model are managed in separate sections of AI Providers.

What's New

• Transcription providers: Save an OpenAI or ElevenLabs API key, test session access, and choose an active transcription model in one searchable custom dropdown.
• OpenAI GPT Live Transcribe: Streams incremental text with the minimal delay setting and commits captured speech turns for final transcripts.
• ElevenLabs Scribe v2 Realtime: Streams partial and committed transcripts through its real-time WebSocket API.
• Independent audio channels: Microphone and meeting audio use separate sessions with the selected provider/model so user and interviewer roles remain distinct.

Improvements

• Renamed the provider-card section to LLM providers while retaining the AI Providers settings tab. The LLM model and OpenAI response tier remain independent of transcription.
• OpenAI transcription uses its separate saved key, or reuses the saved OpenAI LLM key when no separate key is supplied.
• Existing local-only configurations choose an available cloud provider from saved keys. Explicitly selected transcription models persist across restarts and are not silently switched when their key is removed.
• First-run setup now configures cloud transcription instead of downloading a local model. Audio settings link to the transcription section while retaining device controls.
• Connection tests wait for an accepted provider session without recording audio. Transcription changes are disabled during an active interview; stop first, then configure and restart.

Fixes

• Removed forced Moonshine routing and the local model download/preload path.
• Startup audio buffers until the provider accepts the session; buffers are bounded and discarded audio is reported.
• Captured PCM is mixed to mono and converted to the provider sample rate. Small capture chunks are grouped into approximately 100 ms packets.
• OpenAI final results are reconciled by turn identity and emitted in committed-turn order, even when completion events arrive out of order.
• Stop flushes remaining audio and waits up to two seconds for trailing transcript results before saving the interview. Connection failures, rejected credentials, quota errors, and drain timeouts are surfaced.
• Stop cancels reconnect/setup/keepalive timers; stale sockets cannot insert text into a restarted interview. Connection tests clean up immediately after session acceptance.

Technical

• Bumped package version to 2.8.6. Includes the full-file instruction context, Persona, searchable LLM chooser, restartable interviews, DMG signing repairs, and Dock fixes from previous releases.
• Removed local speech model code, the speech worker, model downloader, local-model UI/IPC, Hugging Face Transformers, ONNX Runtime, and unused ElevenLabs SDK/direct gRPC dependencies. No local speech model is bundled or downloaded. Existing user model caches are not automatically deleted.
• Legacy Google/Deepgram SDKs are development dependencies for historical adapter checks and are excluded from production dependencies.
• Full automated suite: 1,053 passed, zero failed, 41 skipped. Renderer type checking and production builds pass. Electron type checking retains the unchanged fork's 37 diagnostics, with no new diagnostics.
• Protocol tests cover accepted sessions, startup buffering, PCM conversion, partial/final results, out-of-order completions, explicit commits, trailing results, bounded buffers, stale sockets, fatal provider errors, and timer cleanup. Credential tests cover upgrade behavior, key reuse, selected-model persistence, and missing keys.
• Browser checks exercise provider/model search, key saving, connection success/error UI, cloud-only onboarding, and start/stop/restart with the same interview identity. Browser checks use simulated Electron IPC; protocol tests use simulated provider sockets.
• Live API audio remains unverified: this environment could not securely pass user API keys into its test runner. Use Test Connection and an installed-Mac audio check to confirm your account's model access, network, transcription accuracy, and latency. No fixed end-to-end latency is promised.
• Audio now streams to the selected cloud transcription provider during interviews. Internet, provider API access, and usage charges are required. Two audio streams may incur separate usage. Updated privacy policy and in-app help reflect this behavior.
• macOS CI verifies app bundles, mounted DMGs, and extracted updater ZIPs before uploading release assets. Native permissions, audio capture, and Dock behavior still require installed-Mac confirmation.

Platform Downloads

macOS

• Apple Silicon: AnswerCue-2.8.6-arm64.dmg or AnswerCue-2.8.6-arm64-mac.zip.
• Intel: AnswerCue-2.8.6-x64.dmg or AnswerCue-2.8.6-mac.zip.
• Auto-update metadata: latest-mac.yml and ZIP blockmaps.

Ad-hoc signed and unnotarized. Replace the previous app in Applications; settings and interview history use the existing data directory. If Gatekeeper blocks this trusted fork build, remove its quarantine flag:

xattr -dr com.apple.quarantine /Applications/AnswerCue.app

Windows

No Windows installer is included in this release.

Linux

No Linux package is included in this release.

AnswerCue v2.8.5

Release date: 2026-10-08

Summary

AnswerCue v2.8.5 keeps chat and repeated interview sessions in one conversation, sends complete instruction files and Persona directly to the LLM, and consolidates model selection into one searchable settings dropdown. It includes the macOS packaging and Dock fixes from v2.8.4.

What's New

• Restartable interviews: Chat to add background, start the interview overlay, stop or close it, continue chatting, and restart the same interview. Earlier chat, transcript, documents, and AI answers remain available.
• Complete instruction-file context: Typed instructions and the full extracted text of the attached MD, TXT, PDF, or DOCX file go directly to the LLM, without embedding or vector search. The counter includes typed text and file content.
• Searchable Active Model: One custom dropdown combines cached chat models from every configured OpenAI, Google Gemini, and Anthropic Claude provider. Search by model or provider.

Improvements

• Fetch Models refreshes and persists that provider's model cache; a failed refresh preserves the previous list. Changing or deleting a provider key invalidates its old cache.
• OpenAI's custom response-tier dropdown offers Auto, Standard, and Fast (Priority). Existing Ultrafast settings migrate to Fast while preserving the selected model.
• Provider cards retain key controls, Test Connection, and Fetch Models. Removed the redundant provider and interview model selectors, and the duplicate interview-sidebar collapse button.
• Pausing and restarting appends new transcript and usage to the same saved interview. Chat stays usable between sessions.
• Production Electron builds include only the main process, preload, speech worker, and PDF worker, without source maps or redundant development bundles.

Fixes

• Saved AI Persona now accompanies Custom Instructions in cloud streaming and other chat requests, including after app restart.
• Removed application-level truncation of attached instruction content and preparation context.
• Closing Settings immediately after typing saves pending instructions and Persona edits. File replacement/removal preserves the latest typed instructions.
• Concurrent starts share one startup result; stop/start waits for teardown. Older background summaries cannot overwrite a newer saved session.
• Supports Responses-only OpenAI Pro/Codex chat models without overriding the chosen model, and expands provider model discovery with paginated Gemini and Claude lists.

Technical

• Bumped package version to 2.8.5; source and release notes are published in this fork.
• Removed MiniLM embedding weights, MobileBERT classifier weights, vector-search modules, and sqlite-vec dependencies. Required local-speech Transformers/ONNX packages remain; Moonshine downloads during setup.
• Full automated suite: 1,049 passed, zero failed, 41 skipped. Renderer type checking and production builds pass. Electron type checking retains the unchanged fork's 37 diagnostics, with no new diagnostics.
• Browser checks exercised cross-provider search, refreshed caches, tier selection, complete-file character counts, immediate-close saving, and chat → start → stop → chat → restart with the same interview ID and earlier context.
• macOS CI verifies app signatures, mounted DMG contents, and extracted ZIP contents before uploading artifacts. Native audio, permissions, and Dock behavior still require confirmation on an installed Mac; browser checks use simulated Electron IPC.
• Complete context increases input tokens and can exceed the selected model's context window. Files above 15 MB are rejected; scanned PDFs require text extraction. The app does not silently trim the instruction file. Priority processing depends on provider/model/account availability and pricing.

Platform Downloads

macOS

• Apple Silicon: AnswerCue-2.8.5-arm64.dmg or AnswerCue-2.8.5-arm64-mac.zip.
• Intel: AnswerCue-2.8.5-x64.dmg or AnswerCue-2.8.5-mac.zip.
• Auto-update metadata: latest-mac.yml and ZIP blockmaps.

Ad-hoc signed and unnotarized. Replace the earlier app in Applications; saved settings and interview history use the existing data directory. If Gatekeeper blocks this trusted fork build, remove its quarantine flag:

xattr -dr com.apple.quarantine /Applications/AnswerCue.app

Windows

No Windows installer is included in this release.

Linux

No Linux package is included in this release.

AnswerCue v2.8.4

Release date: 2026-10-08

Summary

AnswerCue v2.8.4 fixes redundant macOS Dock visibility transitions during startup and reopening, and corrects interview-overlay cleanup during quitting and window recreation.

What's New

• Stable Dock visibility: Reopening reuses the existing Dock entry; concurrent requests share one native visibility transition.

Improvements

• Normal startup keeps the application's existing Dock activation policy.
• Stealth visibility changes converge on the latest setting and respect Electron's hide-after-show cooldown.
• Cancels queued Dock changes during shutdown.

Fixes

• Removed normal startup's accessory-to-regular policy changes.
• Avoids calling dock.show when the icon is already visible.
• A visible interview overlay no longer cancels application quit.
• Closing and recreating the launcher destroys the previous overlay instead of orphaning its renderer.

Technical

• Bumped package version to 2.8.4.
• Eleven focused checks cover normal startup, repeated activation, concurrent show requests, rapid toggles, cooldown cancellation, native errors, shutdown, and overlay cleanup.
• The full suite retains 43 pre-existing failures (895 passed, 41 skipped); Electron type checking has no new diagnostics relative to the unchanged fork.
• The reported duplicate icons appeared with one main app process and correctly configured helper background flags. A Dock restart cleared them temporarily; the code fix removes risky visibility transitions rather than changing helper identities or the single-instance lock.
• macOS package signature verification remains required for app bundles, mounted DMGs, and extracted ZIPs. Installed-Mac confirmation of Dock behavior remains necessary.
• OpenAI response tiers, saved settings, and interview history retain their existing behavior.

Platform Downloads

macOS

• Apple Silicon: AnswerCue-2.8.4-arm64.dmg or AnswerCue-2.8.4-arm64-mac.zip.
• Intel: AnswerCue-2.8.4-x64.dmg or AnswerCue-2.8.4-mac.zip.
• Auto-update metadata: latest-mac.yml and ZIP blockmaps.

Ad-hoc signed and unnotarized. Replace the earlier app in Applications. If Gatekeeper blocks this trusted fork build, remove its quarantine flag:

xattr -dr com.apple.quarantine /Applications/AnswerCue.app

Already-stale Dock entries may need a one-time Dock refresh with killall Dock; then check that restarting AnswerCue does not recreate them.

Windows

No Windows installer is included in this macOS maintenance release.

Linux

No Linux package is included in this release.

AnswerCue v2.8.3

Release date: 2026-10-08

Summary

AnswerCue v2.8.3 fixes macOS packaging defects that can cause the app to be reported as damaged. It retains the OpenAI response-tier settings from v2.8.2.

What's New

• Verified Mac packages: Signature checks cover the built app, the app inside each DMG, and the app extracted from each updater ZIP.

Improvements

• Creates ad-hoc DMGs from pristine app bundles using ditto and hdiutil.
• Preserves nested signatures and native-module entitlements during packaging.
• Release uploads stop when signature, disk-image, or updater-manifest checks fail.

Fixes

• Reseals the outer app after native modules are signed.
• Native-module signing failures now stop the build.
• Avoids electron-builder's signature-changing DMG layout in the ad-hoc packaging path.

Technical

• Bumped package version to 2.8.3.
• Signing regression checks passed locally. Actual macOS package verification runs in GitHub Actions before upload.
• This release is ad-hoc signed, not Developer ID signed or notarized. Valid signatures do not provide automatic Gatekeeper approval; Apple signing/notarization credentials are required for that.
• Manual launch, audio, and permission checks on an installed Mac remain necessary.

Platform Downloads

macOS

• Apple Silicon: AnswerCue-2.8.3-arm64.dmg or AnswerCue-2.8.3-arm64-mac.zip.
• Intel: AnswerCue-2.8.3-x64.dmg or AnswerCue-2.8.3-mac.zip.
• Auto-update metadata: latest-mac.yml and generated blockmaps.

Drag AnswerCue into Applications and replace the earlier app. Saved app data and settings are retained. macOS may require approval in System Settings → Privacy & Security for this trusted fork build. If Gatekeeper still blocks it after copying to Applications, remove the quarantine flag for this app only:

xattr -dr com.apple.quarantine /Applications/AnswerCue.app

This does not turn off Gatekeeper for other apps.

Windows

No Windows installer is included in this macOS packaging-fix release.

Linux

No Linux package is included in this release.

AnswerCue v2.8.2

Release date: 2026-10-08

Summary

AnswerCue v2.8.2 adds a saved OpenAI response-tier selector in Settings and publishes Apple Silicon and Intel DMGs from this fork.

What's New

• Default response tier: Choose Auto, Standard, Fast (Priority), or Ultrafast inside Settings → AI Providers → OpenAI.
• Ultrafast support: Uses GPT 6 Astra through the Responses API for OpenAI text, screenshot, and background answers.
• Mac installers: Architecture-specific DMGs and updater ZIPs are available from this fork.

Improvements

• Saves the tier automatically, applies it to subsequent requests, and restores it after restart.
• Test Connection uses the saved tier.
• Preserves streaming, image inputs, low reasoning, prompt caching, cancellation, and first-token timeouts.
• Update downloads and release notes point to devasheeshG/AnswerCue.

Fixes

• Prevents retries from appending a second answer to a partially streamed response.
• Ultrafast does not silently downgrade to a different OpenAI model.

Technical

• Bumped package version to 2.8.2.
• Auto follows the OpenAI project's default; Standard requests normal processing; Fast requests Priority processing for the selected supported model.
• Ultrafast uses GPT 6 Astra even when another OpenAI model is selected. Higher pricing and lower rate limits apply.
• Transcription, embeddings, other providers, and the separate Groq Fast Response control retain their own settings.
• Renderer/Electron builds and five new transport tests passed. Existing repository test and type-check failures were unchanged; no paid OpenAI requests were made.

Platform Downloads

macOS

• Apple Silicon: AnswerCue-2.8.2-arm64.dmg or AnswerCue-2.8.2-arm64-mac.zip.
• Intel: AnswerCue-2.8.2-x64.dmg or AnswerCue-2.8.2-mac.zip.
• Auto-update metadata: latest-mac.yml and generated blockmaps.

These builds are ad-hoc signed and unnotarized. A packaging defect can cause macOS to report the app as damaged. Use the v2.8.3 packaging-fix release when available.

Windows

No Windows installer was published from this fork for this version.

Linux

No Linux package was published from this fork for this version.