cliptate

Cliptate
On-device AI workspace for Mac audio and video creators
面向 Mac 音视频创作者的端侧 AI 工作台

Website ·
Features ·
Voice & TTS ·
Download

English ·
中文 ·
日本語 ·
한국어 ·
Deutsch

---

Cliptate is a native macOS application that compresses the most time-consuming steps in audio and video production into one continuous, fully local workflow. No cloud upload, no API keys, no waiting on network round trips — every model runs on your Mac.

---

Features

🎙 Voice Dictation — Speak to Type

Real-time streaming speech-to-text that works like an input method. Invoke it from any context with a global keyboard shortcut and get editable text as you speak.

• Streaming transcription with minimal latency
• Automatic multi-speaker diarization and labeling
• Custom vocabulary dictionary for domain-specific accuracy
• Global keyboard shortcut to invoke or pause anywhere

📝 Offline Transcription

Batch-transcribe audio and video files entirely on-device. Supports multiple ASR engines (Parakeet, Qwen3, FireRed ASR) and Apple Silicon-optimized MLX inference for fast, accurate captions without sending audio to any server.

✂️ Smart Pause Trimming

VAD-powered silence detection automatically identifies and removes dead air across podcasts, interviews, and course recordings. Rough cuts arrive closer to delivery-ready with minimal manual scrubbing.

👥 Speaker Diarization

Automatically identify and label multiple speakers across a recording. Sortformer and Silero VAD models run on-device to produce structured, speaker-attributed transcripts for interviews, panels, and meetings.

🔊 TTS Voice Synthesis — Four Modes

| Mode | Description |
|---|---|
| Preset | Built-in multilingual voices, ready instantly |
| Clone | Upload reference audio to reproduce a speaker's timbre and cadence |
| Emotion Control | Steer emotion and speaking style on a named speaker via text prompts |
| Voice Design | Describe a voice in text to synthesize an entirely original persona |

All TTS inference runs on-device via MLX. No audio is sent to the cloud.

🌐 AI Correction & Translation

Context-aware transcript correction followed by on-device or Apple Translation-powered multilingual output. Produce subtitles and scripts in multiple languages without switching tools.

⭐ Highlight Detection

Automatically surface the strongest moments and key sentences from a recording and mark them on the timeline — so publishing edits move faster.

---

Export Formats

Cliptate exports directly into the tools you already use:

| Format | Use |
|---|---|
| FCPXML | Final Cut Pro timeline project exchange |
| SRT | Universal subtitle delivery |
| ASS | Styled subtitles with positioning and formatting |
| VTT | Web captions for browsers and streaming platforms |

Compatible with Final Cut Pro, DaVinci Resolve, and Adobe Premiere Pro.

---

Privacy & Security

• Zero cloud upload — audio, video, and transcripts never leave your Mac
• No API keys required — all models ship with the app or download on first use
• No network dependency — works fully offline once models are cached
• Suitable for sensitive, confidential, and compliance-regulated content

---

Requirements

| | |
|---|---|
| Platform | macOS 14 Sonoma or later |
| Architecture | Apple Silicon (M1 and later) recommended |
| Storage | Varies by model selection (2–8 GB for full model set) |

---

Tech Stack

Cliptate is built entirely in Swift and runs on Apple-native frameworks:

---

Use Cases

Podcasters & Interviewers — Trim silence, generate captions, and flag publishable moments without leaving the app.

Vloggers & Short-form Creators — Smooth out talking-head recordings and keep focus on pacing instead of manual cuts.

Course Creators & Meeting Recorders — Turn hours of live or recorded audio into searchable, translatable text assets via dictation or batch transcription.

Multilingual Content Teams — Clone a voice, translate the transcript, and produce dubbed alternate-language versions — all on one Mac.

---

Links

• Website: cliptate.github.io
• App Store: Download Cliptate
• Issues & Feedback: GitHub Issues

---

Built for creators who care about privacy, speed, and professional output.

Collapse