[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-7117":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":15,"stars7d":16,"stars30d":17,"stars90d":15,"forks30d":15,"starsTrendScore":18,"compositeScore":19,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":22,"hasPages":22,"topics":23,"createdAt":9,"pushedAt":9,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":15,"starSnapshotCount":15,"syncStatus":27,"lastSyncTime":28,"discoverSource":29},7117,"ghost-pepper","matthartman\u002Fghost-pepper","matthartman","100% private on-device voice models for speech-to-text and meeting transcription on macOS",null,"Swift",2770,153,7,46,0,18,183,5,74.56,false,"main",true,[],"2026-06-12 04:00:32","\u003Cdiv align=\"center\">\n\n\u003Cimg src=\".\u002Fapp-icon.png\" width=\"80\" alt=\"Ghost Pepper\">\n\n# Ghost Pepper\n\n**100% private** on-device voice models for speech-to-text and meeting transcription on macOS. No cloud APIs, no data leaves your machine.\n\n\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fmatthartman\u002Fghost-pepper\u002Freleases\u002Flatest\u002Fdownload\u002FGhostPepper.dmg\">\n  \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FDownload_for_Mac-FF6600?style=for-the-badge&logo=apple&logoColor=white\" alt=\"Download for Mac\" height=\"40\">\n\u003C\u002Fa>\n\nmacOS 14.0+ · Apple Silicon (M1+) · Free & open source\n\n[![GitHub stars](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fstars\u002Fmatthartman\u002Fghost-pepper?style=social)](https:\u002F\u002Fgithub.com\u002Fmatthartman\u002Fghost-pepper)\n \n[![License: MIT](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FLicense-MIT-green.svg)](LICENSE)\n \n![100% Local](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002F100%25-Local-FF6600)\n \n![50+ Languages](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002F50%2B-Languages-blue)\n\n\u003C\u002Fdiv>\n\n## Features\n\n- **Hold Control to talk** — release to transcribe and paste into any text field\n- **Meeting transcription** — record calls with notes, transcript, and AI-generated summaries saved as markdown\n- **Runs entirely on your Mac** — models run locally via Apple Silicon, nothing is sent anywhere\n- **Smart cleanup** — local LLM removes filler words and handles self-corrections\n- **Menu bar app** — lives in your menu bar, no dock icon, launches at login\n- **Customizable** — edit the cleanup prompt, pick your mic, toggle features on\u002Foff\n\n## How it works\n\nGhost Pepper uses open-source models that run entirely on your Mac. Models download automatically and are cached locally.\n\n### Speech models\n\n| Model | Size | Best for |\n|---|---|---|\n| Whisper tiny.en | ~75 MB | Fastest, English only |\n| **Whisper small.en** (default) | ~466 MB | Best accuracy, English only |\n| Whisper small (multilingual) | ~466 MB | Multi-language support |\n| Parakeet v3 (25 languages) | ~1.4 GB | Multi-language via [FluidAudio](https:\u002F\u002Fgithub.com\u002FFluidInference\u002FFluidAudio) |\n| Qwen3-ASR 0.6B int8 (50+ languages) | ~900 MB | Highest multilingual quality, macOS 15+ required |\n\n### Cleanup models\n\n| Model | Size | Speed |\n|---|---|---|\n| **Qwen 3.5 0.8B** (default) | ~535 MB | Very fast (~1-2s) |\n| Qwen 3.5 2B | ~1.3 GB | Fast (~4-5s) |\n| Qwen 3.5 4B | ~2.8 GB | Full quality (~5-7s) |\n\nSpeech models powered by [WhisperKit](https:\u002F\u002Fgithub.com\u002Fargmaxinc\u002FWhisperKit). Cleanup models powered by [LLM.swift](https:\u002F\u002Fgithub.com\u002Feastriverlee\u002FLLM.swift). All models served by [Hugging Face](https:\u002F\u002Fhuggingface.co\u002F).\n\n## Getting started\n\n**Download the app:**\n1. Download [GhostPepper.dmg](https:\u002F\u002Fgithub.com\u002Fmatthartman\u002Fghost-pepper\u002Freleases\u002Flatest\u002Fdownload\u002FGhostPepper.dmg)\n2. Open the DMG, drag Ghost Pepper to Applications\n3. Grant Microphone and Accessibility permissions when prompted\n4. Hold Control and speak\n\n> **\"Apple could not verify\" warning?** On macOS Sequoia, you may see a Gatekeeper warning the first time you open the app. Go to **System Settings > Privacy & Security**, scroll down, and click **Open Anyway** next to the Ghost Pepper message. Click **Confirm** in the popup. You only need to do this once.\n\n**Build from source:**\n1. Clone the repo\n2. Open `GhostPepper.xcodeproj` in Xcode\n3. Build and run (Cmd+R)\n\n## Permissions\n\n| Permission | Why |\n|---|---|\n| Microphone | Record your voice |\n| Accessibility | Global hotkey and paste via simulated keystrokes |\n\n## Privacy audit\n\nEvery core feature runs 100% on your Mac — verified by AI code review. No trust required, just point Claude at the repo and ask.\n\n| Feature | Status | What was checked |\n|---|---|---|\n| Speech-to-text | :white_check_mark: Local | WhisperKit\u002FFluidAudio inference, no audio sent anywhere |\n| Text cleanup | :white_check_mark: Local | Qwen LLM runs on-device via LLM.swift |\n| Audio recording | :white_check_mark: Local | AVAudioEngine + ScreenCaptureKit, no streaming |\n| Meeting transcription & storage | :white_check_mark: Local | Chunked transcription, markdown files on disk |\n| Summary generation | :white_check_mark: Local | Local LLM summarization, no cloud API |\n| OCR & screen capture | :white_check_mark: Local | Apple Vision framework, on-device |\n| File storage | :white_check_mark: Local | Markdown to local filesystem, no cloud sync |\n| Analytics & telemetry | :white_check_mark: None | No Firebase, Mixpanel, Sentry, or any tracking SDK |\n\n**Optional cloud features** (disabled by default, require your own API keys): Zo AI chat, Trello integration, Granola meeting import. Model downloads are one-time from Hugging Face.\n\n> **Verify it yourself:** run `cat PRIVACY_AUDIT.md` in Claude Code and ask it to review the codebase against the audit prompt. The [full audit](PRIVACY_AUDIT.md) includes the exact prompt and detailed file-level results.\n\n## Good to know\n\n- **Launch at login** is enabled by default on first run. You can toggle it off in Settings.\n- **Everything stays local** — transcription history and recordings are stored on your Mac only. Nothing is sent to the cloud. You can clear history anytime in Settings.\n\n## Acknowledgments\n\nBuilt with [WhisperKit](https:\u002F\u002Fgithub.com\u002Fargmaxinc\u002FWhisperKit), [LLM.swift](https:\u002F\u002Fgithub.com\u002Feastriverlee\u002FLLM.swift), [Hugging Face](https:\u002F\u002Fhuggingface.co\u002F), and [Sparkle](https:\u002F\u002Fsparkle-project.org\u002F).\n\n## License\n\nMIT\n\n## Why \"Ghost Pepper\"?\n\nAll models run locally, no private data leaves your computer. And it's spicy to offer something for free that other apps have raised $80M to build.\n\n## Enterprise \u002F managed devices\n\nGhost Pepper requires Accessibility permission, which normally needs admin access to grant. On managed devices, IT admins can pre-approve this via an MDM profile (Jamf, Kandji, Mosaic, etc.) using a Privacy Preferences Policy Control (PPPC) payload:\n\n| Field | Value |\n|---|---|\n| Bundle ID | `com.github.matthartman.ghostpepper` |\n| Team ID | `BBVMGXR9AY` |\n| Permission | Accessibility (`com.apple.security.accessibility`) |\n","Ghost Pepper 是一个专为 macOS 设计的语音转文字和会议记录工具，能够在设备上完全私密地运行。其核心功能包括通过按住 Control 键说话并释放来实现即时转录、会议录音及自动生成摘要，并且所有处理都在本地进行，确保数据不会离开用户的设备。该项目利用了如 Whisper 和 Qwen 等先进的开源模型来提供高质量的多语言支持与智能文本清理服务。Ghost Pepper 适用于需要保持高度隐私性的个人或企业用户，在日常办公、在线会议等场景下特别有用，能够有效提高工作效率同时保护敏感信息不被泄露。",2,"2026-06-11 03:10:35","top_language"]