ls-eva

name

ls-eva

description

Use when the user launches Claude from inside the LiteEVA panel, wants to create or work on a video project, or asks about agentic video editing. Subcommands — remotion, voiceover, music, ffmpeg. Triggers on 'LiteEVA', 'open EVA', 'launch EVA', 'video project', 'start a video', 'new video', 'EVA panel', 'video editor', 'create video', 'work on video', 'remotion', 'composition', 'render video', 'voiceover', 'TTS', 'generate narration', 'generate music', 'soundtrack', 'ACE-Step', 'ffmpeg', 'trim video', 'compress video', 'platform export'.

LiteEVA — Agentic Video Editor

You are inside LiteEVA, LiteSuite's agentic video editor. Full video production toolkit — compose, narrate, score, and export without leaving Claude.

Subcommands

Command	What it does
/ls-eva	Start or resume a video project — scaffold structure, open Remotion Studio
/ls-eva remotion	Build, preview, or render Remotion compositions
/ls-eva voiceover	Generate voiceover audio (ElevenLabs or local TTS)
/ls-eva music	Generate background music (ACE-Step, local)
/ls-eva ffmpeg	Post-process: trim, compress, export to platform

---

Toolkit Location

The claude-code-video-toolkit is installed at a configurable path set during LiteEVA's setup wizard. Throughout this skill, this is referred to as [toolkit-dir]. Resolve it to the actual installed path shown in LiteEVA panel settings.

Project Directory Convention

[toolkit-dir]/projects/[project-name]/
  compositions/        # Remotion scene .tsx files
  audio/               # Voiceover + music .mp3/.wav files
  output/              # Rendered video outputs
  frames/              # Extracted frames (for scroll animations)

Port Map

Service	Port	Notes
Remotion Studio	7441	Displayed in LiteEVA's BrowserShellPane
LiteSuite TTS	5123	Local voiceover fallback — no API key required

API Key Management

ElevenLabs API key is set in LiteEVA panel settings and stored securely via electron.safeStorage. Do not write API keys to .env files or hardcode them.

Working Directory Context

The TOOLKIT_CLAUDE_MD_TEMPLATE.md in this skill's directory is the CLAUDE.md template that gets deployed to [toolkit-dir]/CLAUDE.md during setup.

---

Workflow: Start a Video Project

1. Clarify the video — ask what they want to create if not stated
2. Scaffold the project dir:

   mkdir -p [toolkit-dir]/projects/[project-name]/{compositions,audio,output,frames}
   

3. Start Remotion Studio:

   cd [toolkit-dir] && npx remotion studio --port 7441
   

4. Build compositions (/ls-eva remotion)
5. Generate voiceover per scene (/ls-eva voiceover)
6. Generate background music (/ls-eva music)
7. Render: npx remotion render [CompositionId] projects/[name]/output/final.mp4
8. Post-process + export to platform (/ls-eva ffmpeg)

---

Remotion — Video Composition

Remotion turns React components into frame-accurate video. Dev server runs on port 7441 and displays in the BrowserShellPane.

Dev Server

cd [toolkit-dir]
npx remotion studio --port 7441

Shared Component Library

Located at [toolkit-dir]/src/components/:

Component	Description
AnimatedBackground	Full-frame animated background (gradient, particle, wave)
SlideTransition	Slide in/out with configurable direction and easing
Label	Animated text label with fade, typewriter, or pop-in modes
ProgressBar	Frame-synced progress indicator
SceneTitle	Large title card with entrance animation
LowerThird	Broadcast-style lower-third overlay

import { AnimatedBackground, SlideTransition, Label, ProgressBar } from "../components";

Custom Transitions

Available in [toolkit-dir]/src/transitions/:

Transition	Import	Description
glitch	glitchTransition	RGB channel split + scanlines
rgbSplit	rgbSplitTransition	Chromatic aberration sweep
zoomBlur	zoomBlurTransition	Radial blur zoom push
lightLeak	lightLeakTransition	Organic film light leak
clockWipe	clockWipeTransition	Radial clockwise reveal
pixelate	pixelateTransition	Pixelate out/in
checkerboard	checkerboardTransition	Checkerboard reveal

import { glitchTransition } from "../transitions/glitch";
const progress = interpolate(frame, [cutIn, cutIn + 15], [0, 1], { extrapolateRight: "clamp" });
return glitchTransition({ progress, children: <NextScene /> });

Frame Animation Patterns

import { useCurrentFrame, useVideoConfig, interpolate, spring, Sequence } from "remotion";

export const MyComposition: React.FC = () => {
  const frame = useCurrentFrame();
  const { fps, durationInFrames, width, height } = useVideoConfig();
  const opacity = interpolate(frame, [0, 30], [0, 1], { extrapolateRight: "clamp" });
  const scale = spring({ frame: frame - 10, fps, config: { damping: 12, stiffness: 180 } });
  return <div style={{ opacity, transform: `scale(${scale})` }}>Hello Remotion</div>;
};

Sequence Layout

<Sequence from={0} durationInFrames={90}><Intro /></Sequence>
<Sequence from={75} durationInFrames={120}><MainContent /></Sequence>
<Sequence from={180} durationInFrames={60}><Outro /></Sequence>

Composition Definition

Register in [toolkit-dir]/src/Root.tsx:

import { Composition } from "remotion";
import { MyScene } from "./compositions/MyScene";

export const RemotionRoot: React.FC = () => (
  <Composition
    id="MyScene"
    component={MyScene}
    durationInFrames={300}
    fps={30}
    width={1920}
    height={1080}
    defaultProps={{ title: "Demo", accent: "#6366f1" }}
  />
);

Use Case	FPS
Default / export	30
Cinematic	24
Smooth motion	60
GIF export	15

Rendering

cd [toolkit-dir]
npx remotion render MyScene out/my-scene.mp4
npx remotion render MyScene out/my-scene.mp4 --props '{"title":"Episode 1"}'
npx remotion still MyScene out/thumbnail.png --frame 60

Audio Integration

import { Audio, staticFile } from 'remotion';
<Audio src={staticFile('audio/voiceover-scene1.mp3')} startFrom={0} />
<Audio src={staticFile('audio/bg-music.mp3')} volume={0.3} />

Place mp3 files in [toolkit-dir]/public/audio/ for staticFile() resolution.

---

Voiceover — AI Voice Generation

Generates per-scene voiceover using ElevenLabs TTS, with local LiteSuite TTS fallback.

Python Tool

cd [toolkit-dir]

.venv/Scripts/python tools/voiceover.py \
  --text "Welcome to our product demo." \
  --voice JBFqnCBsd6RMkjVDRZzb \
  --output projects/[project-name]/audio/scene1-vo.mp3

# With voice settings
.venv/Scripts/python tools/voiceover.py \
  --text "This is a dramatic line." \
  --voice pNInz6obpgDQGcFmaJgB \
  --stability 0.4 --similarity 0.85 --style 0.6 \
  --output projects/[project-name]/audio/scene2-vo.mp3

Models

Model ID	Use Case
eleven_multilingual_v2	Best quality — default for final output
eleven_flash_v2_5	Fastest — good for drafts
eleven_turbo_v2_5	Balanced

Output Formats

Format	Description
mp3_44100_128	Default, Remotion-compatible
mp3_44100_192	Higher quality
pcm_44100	Lossless

Voice Settings

Parameter	Default	Range	Notes
--stability	0.5	0.0-1.0	Higher = more consistent
--similarity	0.75	0.0-1.0	Higher = closer to original
--style	0.0	0.0-1.0	Style exaggeration

Popular Voices

Name	Voice ID	Character
George	JBFqnCBsd6RMkjVDRZzb	Deep, authoritative British male
Rachel	21m00Tcm4TlvDq8ikWAM	Warm, clear American female
Adam	pNInz6obpgDQGcFmaJgB	Calm, deep American male

SSML Phoneme Control

.venv/Scripts/python tools/voiceover.py \
  --text '<speak>Launch <phoneme alphabet="ipa" ph="ˌdʒuːpɪtər">Jupiter</phoneme> in 3, 2, 1.</speak>' \
  --ssml --voice JBFqnCBsd6RMkjVDRZzb \
  --output projects/[project-name]/audio/scene1-vo.mp3

Local TTS Fallback

Requires LiteSuite running — the TTS service at port 5123 is provided by LiteSuite. If LiteSuite is not running, use ElevenLabs (API key required) or another TTS provider instead.

curl -s -X POST http://127.0.0.1:5123/tts \
  -H "Content-Type: application/json" \
  -d '{"text": "Welcome to the demo.", "voice": "default", "output": "projects/[project-name]/audio/scene1-vo.mp3"}'

Per-Scene Workflow

1. Write scene scripts
2. Generate each voiceover with consistent voice
3. Copy mp3s to [toolkit-dir]/public/audio/
4. Wire into Remotion via <Audio>
5. Adjust timing in Remotion Studio

---

Music — ACE-Step Local Generation

Generates background music using ACE-Step 1.5 — runs entirely on local hardware, no API key required.

Python Tool

cd [toolkit-dir]

.venv/Scripts/python tools/music_gen.py \
  --prompt "corporate ambient electronic, professional, subtle piano, clean, modern, no vocals" \
  --duration 30 \
  --output projects/[project-name]/audio/bg-music.mp3

Task Types

Task	Flag	Description
Text to music	(default)	Generate from prompt
Style transfer	--task cover	Transfer style from reference
Stem extraction	--task stems	Separate vocals, drums, bass, other

Prompt Structure

Format: [Genre] [Emotion] [Instruments] [Timbre] [Era] [Production style] [Vocal style]

"cinematic dramatic, intense, orchestral strings, powerful, epic, no vocals"
"upbeat pop, joyful, acoustic guitar + light percussion, warm, contemporary, lo-fi, no vocals"

Scene Presets

Preset	Prompt
corporate-bg	"corporate ambient electronic, professional, subtle piano, clean, modern, no vocals"
upbeat-tech	"upbeat technology, energetic, synthesizer, bright, contemporary, instrumental"
ambient	"calm ambient, peaceful, pad synths, ethereal, modern, no vocals"
dramatic	"cinematic dramatic, intense, orchestral, powerful, epic, no vocals"

.venv/Scripts/python tools/music_gen.py --preset corporate-bg --duration 30 --output projects/[project-name]/audio/bg.mp3

Duration Guide

Use Case	Duration
Short background loop	15-30s
Scene-length	30-60s
Full intro/outro	60-120s
Feature-length background	120-300s

Volume Layering Convention

Audio Layer	Suggested Volume
Voiceover	1.0 (full)
Background music	0.2-0.35
Sound effects	0.5-0.75

---

FFmpeg — Post-Processing & Platform Export

FFmpeg handles all non-Remotion video work: conversion, trimming, platform exports, post-processing.

Binary

Use ffmpeg from your system PATH or from the bundled binary if LiteSuite is installed:

# If LiteSuite is installed, use the bundled binary (requires LiteSuite running):
FFMPEG="${LITESUITE_DIR:-<your-litesuite-dir>}/tools/ffmpeg/ffmpeg.exe"

# Otherwise, use system ffmpeg (install: https://ffmpeg.org/download.html or `winget install ffmpeg`):
FFMPEG="ffmpeg"

Common Operations

GIF to MP4

$FFMPEG -i input.gif -vf "scale=trunc(iw/2)*2:trunc(ih/2)*2" -movflags faststart -pix_fmt yuv420p -vcodec libx264 -crf 23 output.mp4

Resize

$FFMPEG -i input.mp4 -vf "scale=1920:1080:force_original_aspect_ratio=decrease,pad=1920:1080:(ow-iw)/2:(oh-ih)/2" output_1080p.mp4

Compress

$FFMPEG -i input.mp4 -vcodec libx264 -crf 24 -preset slow -movflags faststart output.mp4

Trim / Cut

$FFMPEG -ss 00:00:10 -to 00:01:20 -i input.mp4 -c copy trimmed.mp4

Speed Adjustment

$FFMPEG -i input.mp4 -filter_complex "[0:v]setpts=PTS/2.0[v];[0:a]atempo=2.0[a]" -map "[v]" -map "[a]" output_2x.mp4

Extract Audio

$FFMPEG -i input.mp4 -vn -acodec libmp3lame -q:a 2 audio.mp3

Concatenation

cat > concat_list.txt << 'EOF'
file 'clip1.mp4'
file 'clip2.mp4'
EOF
$FFMPEG -f concat -safe 0 -i concat_list.txt -c copy joined.mp4

Fade In / Fade Out

$FFMPEG -i input.mp4 -vf "fade=t=in:st=0:d=1,fade=t=out:st=9:d=1" -af "afade=t=in:st=0:d=1,afade=t=out:st=9:d=1" output_fades.mp4

Silence Removal

$FFMPEG -i input.mp4 -af "silenceremove=stop_periods=-1:stop_duration=0.5:stop_threshold=-50dB" output_no_silence.mp4

Captions (SRT)

$FFMPEG -i input.mp4 -vf "subtitles=captions.srt:force_style='FontSize=24,PrimaryColour=&Hffffff'" output_captions.mp4

Color Grading

$FFMPEG -i input.mp4 -vf "eq=brightness=0.05:contrast=1.1:saturation=1.2" output_graded.mp4

Platform Presets

YouTube (1080p60)

$FFMPEG -i input.mp4 -vf "scale=1920:1080:force_original_aspect_ratio=decrease,pad=1920:1080:(ow-iw)/2:(oh-ih)/2" -vcodec libx264 -crf 18 -preset slow -r 60 -acodec aac -b:a 192k -movflags faststart output_youtube.mp4

TikTok (9:16 vertical)

$FFMPEG -i input.mp4 -vf "scale=1080:1920:force_original_aspect_ratio=decrease,pad=1080:1920:(ow-iw)/2:(oh-ih)/2" -vcodec libx264 -crf 23 -preset slow -acodec aac -b:a 128k -movflags faststart output_tiktok.mp4

Instagram (Square 1:1)

$FFMPEG -i input.mp4 -vf "crop=min(iw\,ih):min(iw\,ih),scale=1080:1080" -vcodec libx264 -crf 23 -preset slow -acodec aac -b:a 128k -movflags faststart output_instagram.mp4

Twitter/X (1080p, max 2:20)

$FFMPEG -i input.mp4 -vf "scale=1920:1080:force_original_aspect_ratio=decrease,pad=1920:1080:(ow-iw)/2:(oh-ih)/2" -vcodec libx264 -crf 23 -preset slow -acodec aac -b:a 128k -movflags faststart -t 00:02:20 output_twitter.mp4

LinkedIn (720p+)

$FFMPEG -i input.mp4 -vf "scale=1280:720:force_original_aspect_ratio=decrease,pad=1280:720:(ow-iw)/2:(oh-ih)/2" -vcodec libx264 -crf 23 -preset slow -acodec aac -b:a 128k -movflags faststart output_linkedin.mp4

WebM (VP9)

$FFMPEG -i input.mp4 -vcodec libvpx-vp9 -crf 33 -b:v 0 -acodec libopus -b:a 128k output.webm

Frame Extraction (for scroll animations)

mkdir -p [toolkit-dir]/projects/[project-name]/frames
$FFMPEG -i rendered.mp4 -vf "fps=30" -q:v 2 [toolkit-dir]/projects/[project-name]/frames/frame_%04d.jpg