Narrator

AI-powered video narration generator. Import a video, configure your narration style, and let AI generate timed narration scripts with optional text-to-speech audio.

Features

Narrator follows a 6-step wizard workflow:

1. Project Setup -- Import video files or record your screen directly
2. Edit Video -- Trim, split, reorder clips, adjust speed, and drop frames
3. Configuration -- Choose AI provider, narration style, language, and attach context documents
4. Processing -- Frame extraction, document processing, and AI narration generation
5. Review -- Preview the generated narration script with segment-level editing
6. Export -- Export scripts as JSON, SRT, VTT, TXT, Markdown, or SSML; generate TTS audio via ElevenLabs

Tech Stack

• Desktop framework: Tauri v2 (Rust backend + webview frontend)
• Backend: Rust (async with Tokio)
• Frontend: React 19, TypeScript, Tailwind CSS, Zustand
• Video processing: ffmpeg / ffprobe
• AI providers: Claude (Anthropic) and OpenAI
• Text-to-speech: ElevenLabs

Installation

Download the latest release from GitHub Releases.

macOS

The app is not code-signed, so macOS will block it by default. After downloading the .dmg:

1. Open the .dmg and drag Narrator.app to /Applications
2. Open Terminal and run:

   sudo xattr -rd com.apple.quarantine /Applications/Narrator.app

3. Open Narrator from Applications

Windows

Windows SmartScreen may show a "Windows protected your PC" warning. Click More info → Run anyway to proceed.

Prerequisites (Development)

• Rust (stable toolchain)
• Node.js (v18+) with pnpm
• curl and unzip (for the ffmpeg fetch script). The app bundles its own ffmpeg, so you don't need one on PATH.

Getting Started

# Install frontend dependencies
pnpm install

# One-time: download the self-contained ffmpeg sidecars (with libass).
# The default Homebrew/apt builds ship without libass, which silently
# breaks burn-subtitles export, so the project uses curated builds instead.
./scripts/fetch-ffmpeg.sh

# Run in development mode
pnpm tauri dev

# Build for production
pnpm tauri build

Project Structure

VideoNarrator/
  src/                      # React frontend
    components/             # Reusable UI components
    features/               # Feature screens (wizard steps)
    hooks/                  # Custom React hooks
    lib/                    # Utilities and Tauri command bindings
    stores/                 # Zustand state stores
    types/                  # TypeScript type definitions
  src-tauri/                # Rust backend
    src/
      lib.rs                # Tauri app entry point
      commands.rs           # Tauri command handlers
      models.rs             # Shared data models
      ai_client.rs          # Claude and OpenAI integration
      video_engine.rs       # ffmpeg frame extraction and probing
      video_edit.rs         # Video editing operations
      doc_processor.rs      # Document context extraction
      export_engine.rs      # Script export formatters
      project_store.rs      # Project persistence
      elevenlabs_client.rs  # ElevenLabs TTS client
      screen_recorder.rs    # Native screen recording
      error.rs              # Error types

Using Narrator with Claude Code

Narrator ships a Claude Code skill that wraps the narrator-cli binary. Once installed, you can ask Claude Code to narrate a video, burn subtitles, probe media, or run small edits conversationally.

# 1. Symlink the skill into your Claude Code skills directory.
#    `$(pwd)` resolves to your current directory — run this from the
#    Narrator repo root (`cd` into it first) so the link points at the
#    checked-out skill.
cd /path/to/VideoNarator
mkdir -p ~/.claude/skills
ln -s "$(pwd)/skills/narrator" ~/.claude/skills/narrator

# 2. Put narrator-cli on PATH (pick whichever fits your setup):
#    a) Build from source — ends up in ~/.cargo/bin, no sudo:
cargo install --path src-tauri --bin narrator-cli
#    b) Symlink into a user-writable PATH dir (no sudo):
mkdir -p ~/.local/bin
ln -sf /Applications/Narrator.app/Contents/MacOS/narrator-cli ~/.local/bin/narrator-cli
#    c) System-wide (asks for sudo):
# sudo ln -sf /Applications/Narrator.app/Contents/MacOS/narrator-cli /usr/local/bin/narrator-cli

# 3. Verify everything is wired up
./skills/narrator/scripts/doctor.sh

See skills/narrator/SKILL.md for the skill's capabilities and skills/narrator/references/install.md for platform-specific install notes.

License

MIT