faceless-script

name	faceless-script
description	Generate faceless TikTok voiceover scripts from repo content for The Terminal Alchemist character. Outputs character dialogue + expression cues + screen recording notes in 15-30 seconds. Use when the user types /faceless or asks to create a faceless TikTok script.

Faceless Script — Terminal Alchemist Voiceover Generator

Generates a voiceover script for The Terminal Alchemist faceless TikTok account. Pulls from repo content (skills, plays, tips, builds) and outputs character dialogue with expression cues, ElevenLabs-ready text, and screen recording B-roll notes. 15-30 seconds per video.

The key difference from /tiktokscript: these scripts are written as character voiceover dialogue with expression cues for the animated alchemist, not on-screen text scripts for face-to-camera recording.

Command Pattern

• /faceless — generate a script from current conversation context
• /faceless <topic> — generate a script for a specific topic
• /faceless repurpose <path> — compress a LinkedIn/X post into a faceless voiceover script
• /faceless batch — generate 3-5 scripts from queued topics
• Also triggers when the user says: "faceless script for...", "alchemist video about...", "terminal alchemist script"

When to Use

• User wants a voiceover script for the faceless TikTok account
• User wants to repurpose LinkedIn/X content into animated character videos
• User wants to batch-generate scripts from the repo content mine
• User has a tool tip, build moment, or pattern to turn into a character monologue

Workflow

Step 1: Identify the Source

1. Topic provided: Use the user's topic directly
2. Repurpose path provided: Read the source file and extract the key moment
3. Batch mode: Pull from queued topics in workflows/tiktok-index.md and content mine mapping
4. No source: Ask: "What's the topic? Tool tip, build moment, or pattern — one sentence."

Step 2: Load References

Read before scripting:

• skills/tier-2-context-playbooks/tiktok.md — TikTok playbook (structure, voice, series)
• skills/tier-1-voice-dna/core-voice.md — core voice principles
• skills/tier-1-voice-dna/viral-hooks.md — hook styles (compressed for 2-second video hooks)
• content/tiktok/faceless/character-decision.md — character spec (expressions, palette, identity)
• content/tiktok/faceless/voice-profile.md — ElevenLabs voice selection, settings, and pacing reference
• workflows/tiktok-index.md — check active series and episode numbers

If repurposing, also read the source file.

Step 3: Determine Series Fit

Match the topic to a faceless series:

Series	Best For
Terminal Alchemist Tips	General tool tips, shortcuts, workflow moments
Skills That Run My Business	Deep dive on one repo skill per video
30-Second GTM Plays	Repurposed from LinkedIn plays series
Claude Code Secrets	Claude Code-specific tricks and commands
Cursor Tricks You Missed	Cursor IDE features and MCP integrations
How to Not Sound Like AI	Voice system, anti-slop, content quality

If no series fits, script as standalone.

Step 4: Generate the Script

File path: content/tiktok/faceless/drafts/YYYY-MM-DD_{slug}.md

File structure:

# {Title}

> **Platform**: TikTok / Reels / Shorts (Faceless)
> **Series**: {Series name or "Standalone"}
> **Episode**: {# if part of a series}
> **Character**: The Terminal Alchemist
> **Source**: {Path to source file if repurposed, or "Original"}
> **Date**: YYYY-MM-DD
> **Status**: draft
> **Est. Duration**: {X seconds}

---

## Hook (0-3 seconds)

**On-screen text**: "{bold text overlay — 8 words or fewer}"

**Voiceover**:
> {Character's opening line. Conversational, confident, hooks immediately.}

**Expression**: {neutral | explaining | reacting | confident | thinking}

**Visual**: {What's behind the character — dark bg, terminal text, or screen recording starting}

---

## Body (3-20 seconds)

**Voiceover**:
> {Main dialogue. 2-4 sentences max. Each sentence is a beat.}
> {The demo or insight — what the alchemist is teaching.}
> {The result or payoff.}

**Expression sequence**:
1. {expression} — "{corresponding line}"
2. {expression} — "{corresponding line}"
3. {expression} — "{corresponding line}"

**Screen recording / B-roll**:
- {What to show behind the character during each beat}
- {Zoom, crop, highlight notes}

**On-screen text overlays**:
- {Key phrases to display as text during the voiceover — 2-4 words each}

---

## Closer (20-30 seconds)

**Voiceover**:
> {Closing line. Soft CTA or loop setup.}

**Expression**: {confident | thinking}

**On-screen text**: "{CTA text — 'save this' / 'follow for more' / 'part 2?'}"

**Visual loop**: {Does end connect to start? Describe the loop if applicable.}

---

## ElevenLabs Text Block

```text
{Clean voiceover text — no stage directions, no formatting.
Just the words the character says, paragraph-separated by beat.
This block gets pasted directly into ElevenLabs.}

---

Caption

{1-2 line caption, lowercase first word, capitalize I, soft CTA}

Tags

{3-5 relevant hashtags}

---

Production Notes

Screen recording needed: {Yes/No — what to record} B-roll alternative: {Gameplay, terminal animation, or pre-recorded clip from library} Audio: ElevenLabs TTS → Chris voice (see content/tiktok/faceless/voice-profile.md for settings) Lip sync: SadTalker / D-ID (use neutral expression as base) Compositing: Character lower-right, screen recording as main content Cross-post to: Reels, Shorts{, LinkedIn video if GTM-relevant}

---

Notes

{Source material reference, key data points, why this topic works for the character}

### Step 5: Write the Voiceover Dialogue

The Terminal Alchemist's voice follows the same core-voice.md principles but adapted for spoken word:

**Character voice rules**:
- First person. Builder who does the work.
- Casual competence. Not lecture-y. Not tutorial-channel.
- Short sentences. Each sentence is a beat.
- Pattern recognition energy. Name what others haven't seen.
- Conversational pacing. Imagine explaining to someone over your shoulder.
- No "hey guys." No "today we're going to." Just start.
- Contractions always (don't, can't, won't, I'm, I've).
- Technical terms are fine but explain by showing, not defining.

**Dialogue pacing for spoken word** (calibrated to Chris voice, see `content/tiktok/faceless/voice-profile.md`):
- 15-second video: ~35-42 words
- 20-second video: ~50-58 words
- 25-second video: ~62-72 words
- 30-second video: ~75-85 words
- Each sentence gets its own beat and expression cue

### Step 6: Generate Hook Options

Produce **3 hook options** for the on-screen text, using `viral-hooks.md` TikTok rules:

- 8 words or fewer
- Must work without sound
- Lowercase first word (unless I)
- Creates enough tension to stop the scroll in 1 second

### Step 7: Save and Confirm

1. **Save script**: `content/tiktok/faceless/drafts/YYYY-MM-DD_{slug}.md`
2. **Display confirmation**:

Faceless script drafted:

content/tiktok/faceless/drafts/YYYY-MM-DD_{slug}.md - Series: {series name or Standalone} - Duration: ~{X} seconds - Word count: {N} words (voiceover) - Expression cues: {list}

Next steps:

• Review voiceover text
• Generate TTS via ElevenLabs
• Run lip sync on character base image
• Record screen capture for B-roll
• Composite and post

## Content Mine — Repo Source Mapping

Your repo already has 150+ videos worth of material:

| Repo Source | Faceless Series | Est. Videos |
|-------------|----------------|-------------|
| 44 agent skills | Skills That Run My Business | 44+ |
| GTM Plays (LinkedIn) | 30-Second GTM Plays | 20+ |
| X micro-tips | Terminal Alchemist Tips | 30+ |
| Voice system / anti-slop | How to Not Sound Like AI | 10+ |
| Partner workflows (anonymized) | 30-Second GTM Plays | 15+ |
| Claude Code tricks | Claude Code Secrets | 25+ |
| Cursor skills | Cursor Tricks You Missed | 20+ |

## Repurpose Logic

When the source is a LinkedIn play or X thread:

1. **Read the source** — find the single most visual, most impactful moment
2. **Extract the character's take** — what would the alchemist say about this? Not a summary. A reaction + lesson.
3. **One insight per video** — strip everything else. One moment, one lesson.
4. **Hook with the result** — start with what happened, then show how
5. **Write it conversational** — this is spoken word, not written content

## Batch Mode

When `/faceless batch` is called:

1. Read `workflows/tiktok-index.md` for queued topics
2. Scan `content/linkedin/drafts/` and `content/x/drafts/` for high-potential repurpose targets
3. Generate 3-5 scripts covering different series
4. Save all to `content/tiktok/faceless/drafts/`
5. Present a summary table of all generated scripts

## Error Handling

- **No topic provided**: "What's the topic? Tool tip, build moment, or pattern. One sentence."
- **Topic too complex for 30 seconds**: "That's a full tutorial. What's the ONE moment that makes someone say 'wait, how?' Let's script around just that."
- **Source file not found**: "Couldn't find that file. Drop the path or describe the tip."
- **No matching series**: Script as standalone. Suggest which series it could start.

## Collaborates With

- **TikTok Script** (`tiktok-script/SKILL.md`): Face-to-camera version. Faceless scripts follow the same structure but add character dialogue and expression cues.
- **Viral Hooks** (`viral-hooks/SKILL.md`): Hook options compressed to 8 words for on-screen text
- **Play Draft** (`play-draft/SKILL.md`): LinkedIn/X drafts feed `/faceless repurpose`
- **Content Images** (`content-images/SKILL.md`): Can generate thumbnail graphics for the video
- **ElevenLabs**: TTS generation from the ElevenLabs Text Block section