Build Wealth Not Just a Paycheck Podcast Thumbnail

Create a **premium-quality, ultra-realistic YouTube podcast/interview thumbnail** inspired by the uploaded reference image, matching its **overall composition, two-person portrait placement, microphone setup, typography hierarchy, warm orange/black color palette, dramatic studio lighting, depth, contrast, conversational posing, and polished high-CTR YouTube-thumbnail aesthetic**.

# IMPORTANT — Ask for User Inputs First

Because this thumbnail design contains **two different people**, before generating the final thumbnail, ask the end user in **ONE message** for ALL THREE required inputs:

1. **Upload the first person's photo** to use as the facial identity reference for the person on the **left side**.
2. **Upload the second person's photo** to use as the facial identity reference for the person on the **right side**.
3. **Tell me the exact thumbnail/content title you want displayed.**

Use this exact wording:

**“Please upload two photos — one for the person on the left and one for the person on the right — and tell me the exact title/headline you want displayed in this thumbnail.”**

Do **not** generate the final thumbnail until:

- Photo 1 is provided
- Photo 2 is provided
- the exact title/headline is provided

Do not use one person's identity for both characters unless the user explicitly requests it.

If only one photo is provided, **do not generate the thumbnail yet**. Ask for the second person's photo.

---

# CRITICAL — UPLOADED PHOTOS ARE IDENTITY-ONLY REFERENCES

The two uploaded photos control **facial identity only**.

They must **NOT control the pose, body position, head angle, shoulder angle, gaze direction, expression, crop, clothing, background, lighting, or camera angle**.

The **reference thumbnail must control the posing and conversational body language**.

This rule applies equally to **BOTH people**.

If either uploaded photo shows a pose similar to what the AI would naturally reproduce, **do not simply copy that pose**.

Instead, intentionally reconstruct the person's body language so it matches the **reference thumbnail's interview interaction** while still looking natural.

## VERY IMPORTANT POSE VARIATION RULE

The two people must **NOT look like two independent portraits pasted beside each other**.

They must look like they are **actively participating in the same podcast conversation**.

Their poses should be **complementary but slightly different**:

- different torso angles
- different head rotations
- slightly different shoulder levels
- different facial expressions
- different mouth positions
- different gaze directions
- natural asymmetry

If one person's generated pose accidentally resembles their uploaded input photo too closely, **change the pose noticeably while preserving their identity**.

The final pose must be driven by the **reference thumbnail**, not by the identity photos.

---

# IMPORTANT — Reference Text Is Style-Only

The wording visible in the reference image, including **“TEACHER,” “$250K,” and “IN 2 YEARS,” must NOT automatically be used as the final thumbnail wording.**

Treat that wording only as a guide for:

- typography size
- placement
- hierarchy
- line breaks
- white/orange color treatment
- black shadow/stroke
- dimensional text effect
- emphasis on the main numeric/keyword phrase

The **user's exact supplied title must control the final content**.

Do not infer, paraphrase, shorten, expand, rewrite, or invent additional readable wording.

# STRICT TEXT RULE

**The only readable text permitted anywhere in the finished thumbnail is the exact title/headline supplied by the user.**

Do not add:

- names
- subtitles
- labels
- channel names
- usernames
- descriptions
- badges
- extra numbers
- decorative pseudo-text

---

# Canvas & Composition

Create a **16:9 horizontal YouTube thumbnail**, approximately **1280 × 720 px**.

Closely reproduce the reference composition:

- first person occupying the **far left**
- second person occupying the **far right**
- both shown as large chest-up podcast/interview portraits
- large headline centered between and slightly behind the two people
- microphones extending inward from the lower-left and lower-right areas
- dark atmospheric podcast/studio environment
- brightest visual emphasis concentrated around the people and headline
- balanced composition with natural conversational interaction

Each person should occupy approximately **30–35% of the frame width**, leaving enough central space for the headline.

Do not make either person unnecessarily small.

---

# LEFT PERSON — PHOTO 1

Use the **first uploaded photo as the only facial identity reference for the left-side person**.

Position this person near the **far-left edge**, framed approximately from the upper chest upward.

## LEFT PERSON — REQUIRED REFERENCE-BASED POSE

Do **not copy Photo 1's original pose**.

Reconstruct the body position according to the reference thumbnail.

Use approximately:

- torso turned slightly inward toward the center
- one shoulder subtly closer to the camera than the other
- upper body leaning very slightly toward the conversation
- head gently rotated toward the right-side participant
- chin positioned naturally rather than perfectly centered
- eyes directed toward the other speaker
- relaxed shoulders
- conversational, engaged posture
- warm, attentive expression
- subtle smile or mild speaking reaction
- mouth slightly open if appropriate

Avoid a rigid front-facing portrait.

Avoid perfectly symmetrical shoulders.

The person should feel like they are **listening and reacting naturally during a real podcast conversation**.

---

# RIGHT PERSON — PHOTO 2

Use the **second uploaded photo as the only facial identity reference for the right-side person**.

Position this person near the **far-right edge**, framed approximately from the upper chest upward.

The right-side person may appear slightly larger or somewhat more visually dominant, similar to the reference.

## RIGHT PERSON — REQUIRED POSE CHANGE

This is especially important:

**Do NOT reproduce the pose, shoulder position, head angle, gaze, body angle, or expression from Photo 2.**

Even if Photo 2 contains a usable podcast-style pose, treat it as **identity information only**.

Rebuild the body language to follow the reference thumbnail.

The right-side pose should be **visibly different from the uploaded input-photo pose**.

Use approximately:

- torso angled inward toward the left-side person
- body rotated slightly rather than facing the camera perfectly straight
- one shoulder slightly forward
- upper torso with a subtle natural lean
- head rotated toward the left participant
- chin angle changed naturally from the source photo
- eyes focused toward the other speaker rather than directly at camera
- energetic speaking expression
- mouth naturally open mid-sentence
- expressive eyes
- eyebrows slightly raised if natural
- lively but professional host/guest presence
- optional subtle hand gesture entering near the lower-right area

The right person's pose should feel like a **captured moment during an active conversation**, not a static portrait.

### MANDATORY POSE DIFFERENCE

If Photo 2 shows:

- straight-facing shoulders → rotate the generated shoulders inward
- camera-facing gaze → redirect the gaze toward the other participant
- closed mouth → use a natural mid-speaking expression
- upright rigid posture → introduce a subtle conversational lean
- symmetrical shoulders → create slight asymmetry
- head facing straight → rotate the head subtly toward the left

Do this while preserving Photo 2's facial identity accurately.

---

# TWO-PERSON INTERACTION — HIGH PRIORITY

The two people should look like they were photographed **together during the same interview**, even though their identities come from separate uploaded photos.

Create believable conversational interaction.

The left person should appear more like they are:

- listening
- reacting
- smiling subtly
- preparing to respond

The right person should appear more like they are:

- actively explaining
- speaking
- emphasizing a point
- engaging the other participant

Do not make both people use the same:

- head angle
- expression
- shoulder angle
- mouth position
- eye direction
- body lean

Their differences should create the natural visual rhythm seen in a real podcast interview.

---

# FACIAL IDENTITY — HIGHEST PRIORITY

Each uploaded photo belongs to a **different person**.

Never mix their identities.

### LEFT PERSON

Photo 1 controls only the left person's identity.

### RIGHT PERSON

Photo 2 controls only the right person's identity.

For each person preserve accurately:

- face shape
- facial proportions
- forehead
- natural hairline
- hairstyle
- hair texture
- hair volume
- eyebrows
- eye shape
- eye spacing
- nose shape
- lips
- mouth shape
- cheeks
- jawline
- chin
- ears where visible
- skin tone
- apparent age
- facial hair if present
- recognizable identity

Do not:

- beautify the faces
- reshape facial features
- slim the face
- change ethnicity
- significantly change age
- over-smooth skin
- create generic AI faces
- merge features from the two uploaded people

Avoid:

- face-swap appearance
- pasted-on face
- floating head
- mismatched neck and body
- incorrect skin-tone transitions
- distorted facial proportions

The head, neck, shoulders, and visible body of each person must appear naturally integrated as the **same individual**.

**Identity must remain accurate even when the pose is changed significantly from the uploaded identity photo.**

---

# MICROPHONES

Include exactly **two professional black podcast microphones**.

### Left Microphone

Place a large black broadcast-style microphone in the lower-left / center-left foreground.

It should:

- extend diagonally toward the left person's mouth
- include a believable boom arm
- include a realistic mount or shock mount
- appear properly positioned for podcast recording

### Right Microphone

Place the second black broadcast microphone near the lower-right / center-right foreground.

It should:

- extend diagonally toward the right person's mouth
- use realistic professional podcast hardware
- remain visually separate from the left microphone

Do not add additional microphones.

The microphone positioning should reinforce the altered conversational poses rather than forcing the people into their original input-photo posture.

---

# Clothing

Use **professional, premium casual podcast/interview clothing**.

Suitable colors:

- black
- charcoal
- dark navy
- deep brown
- muted dark neutrals

Avoid:

- loud patterns
- large logos
- readable clothing text
- distracting accessories

Do not automatically copy the clothing visible in either uploaded identity photo.

The clothing should naturally fit the newly reconstructed body pose.

---

# HEADLINE TYPOGRAPHY

Place the exact user-provided title in the **upper-middle / center area between the two people**.

The headline should be the strongest graphic element after the faces.

Use a:

- very heavy
- condensed
- uppercase
- premium sans-serif display font

similar to high-performing YouTube business/podcast thumbnails.

Adapt the line breaks intelligently according to the length of the supplied title.

The user's exact wording must remain unchanged.

Do not remove or substitute words just to improve the layout.

---

# TYPOGRAPHY HIERARCHY

Break the supplied title visually into approximately **2–3 levels** where appropriate.

### Primary Wording

Use:

- bright white
- extremely bold lettering
- thick black outline
- deep dark drop shadow
- subtle dimensional/extruded effect

### Most Important Word, Number or Phrase

Identify the most visually important portion of the user's exact title **without changing its wording**.

Make this phrase:

- substantially larger
- bright white
- extremely bold
- strongly shadowed
- positioned over a vivid orange backing panel

### Supporting Title Words

Use:

- white and/or orange treatment
- smaller size than the dominant phrase
- strong dark shadow
- excellent readability at small thumbnail size

Do not create additional wording.

---

# ORANGE HEADLINE PANEL

Place a **vivid orange / amber rectangular or slightly skewed parallelogram panel** behind the dominant title phrase.

The shape should:

- extend slightly beyond the letters
- have a subtle tilt or perspective
- feel dimensional
- separate the headline from the dark background
- resemble the visual impact of the reference thumbnail

Use:

- burnt orange
- vivid amber
- warm orange

Avoid neon orange.

---

# TYPOGRAPHY DEPTH

Create layered depth similar to the reference.

Some headline portions may appear **behind the foreground portraits**, while key wording remains clearly readable.

Use natural overlap so that the people feel positioned in front of the typography.

Do not obscure important title words excessively.

---

# BACKGROUND

Create a **dark cinematic podcast studio environment**.

Use predominantly:

- deep black
- charcoal
- dark brown
- burnt orange
- amber
- bronze

Keep the environment intentionally blurred and atmospheric.

Possible subtle details:

- dark acoustic panels
- studio curtains
- soft amber practical lights
- orange light leaks
- subtle smoke or haze
- floating dust particles
- faint glowing specks
- warm background bokeh

Keep the central area behind the headline relatively dark for readability.

Do not add:

- signs
- posters
- logos
- screens with readable content
- additional people
- random text

---

# LIGHTING

Use premium professional podcast-studio lighting.

## Both Faces

Apply:

- soft neutral key lighting
- realistic facial highlights
- detailed natural skin texture
- bright but believable eye catchlights
- strong subject/background separation
- controlled shadow falloff

Add **warm orange/amber rim lighting** from behind each subject.

The outer edges of:

- hair
- shoulders
- face profile

should catch a subtle golden-orange glow.

The rim lighting should enhance separation without creating artificial halos.

Lighting must adapt naturally to the **new reference-based poses**, not the uploaded source-photo lighting.

---

# COLOR PALETTE

Maintain a strong cinematic combination of:

- black
- charcoal
- dark brown
- burnt orange
- amber
- warm gold
- bright white

Do not introduce unrelated saturated colors.

Maintain a cohesive **orange + black premium podcast brand aesthetic**.

---

# DEPTH & LAYERING

Build the thumbnail using these visual layers:

1. dark blurred podcast studio background
2. warm amber lights, haze, and particles
3. central orange headline panel
4. oversized exact user-provided title
5. left portrait using Photo 1 identity with reconstructed reference-based pose
6. right portrait using Photo 2 identity with reconstructed reference-based pose
7. foreground microphones

Create believable separation between every layer.

---

# TITLE-BASED VISUAL ADAPTATION

The exact user-provided title should control the **topic and visual emphasis** of the thumbnail.

The title may concern:

- business
- career
- entrepreneurship
- finance
- education
- marketing
- technology
- AI
- success stories
- productivity
- interviews
- personal growth

Adapt the mood and emphasis appropriately while keeping the reference's podcast composition intact.

However:

**Do not add topic-related icons, logos, symbols, screenshots, labels, or extra visual text unless explicitly requested by the user.**

The main adaptation should come through:

- headline hierarchy
- expressions
- posing
- conversational interaction
- lighting emphasis
- atmosphere
- visual storytelling

---

# IMAGE QUALITY

Render as:

- ultra-realistic photography
- professional podcast photography
- premium commercial YouTube thumbnail
- razor-sharp facial details
- realistic hair strands
- natural skin pores
- accurate eyes and teeth
- realistic hands
- accurate fingers
- realistic professional microphones
- cinematic studio lighting
- shallow depth of field
- HDR-style contrast
- crisp title typography
- polished commercial retouching
- strong subject separation
- high clarity at small thumbnail size

Both people must remain clearly recognizable as the identities supplied in Photo 1 and Photo 2.

---

# NEGATIVE INSTRUCTIONS

Do not include:

- the reference headline automatically
- “TEACHER”
- “$250K”
- “IN 2 YEARS”
- additional captions
- invented subtitles
- names unless literally part of the supplied title
- usernames
- channel names
- logos
- watermarks
- random UI elements
- readable background text
- pseudo-text
- duplicated people
- cloned identities
- identity blending
- swapped identities
- extra microphones
- malformed hands
- extra fingers
- warped faces
- plastic skin
- pasted-face effects
- mismatched head and body
- excessive orange saturation
- clutter around the headline
- generic AI facial replacements

## POSE-RELATED NEGATIVE INSTRUCTIONS

Do not:

- copy Photo 1's pose
- copy Photo 2's pose
- preserve the original body angle from either uploaded photo
- keep the original shoulder orientation from either uploaded photo
- reproduce the original gaze direction
- use the source photo's original crop
- make both people face directly toward the camera
- make both people use identical poses
- make both people use identical expressions
- make both people perfectly upright
- make both shoulders perfectly symmetrical
- create passport-photo body language
- create two unrelated portrait cutouts
- create stiff mannequin-like posing

---

# FINAL NON-NEGOTIABLE REQUIREMENT

The completed thumbnail must preserve the reference design's:

- **two-person podcast/interview composition**
- **reference-inspired conversational body language**
- **clearly different but complementary poses for the two participants**
- **two professional microphones**
- **large central typography**
- **orange title backing panel**
- **dark orange/black podcast environment**
- **warm rim lighting**
- **dramatic contrast**
- **layered subject/text depth**
- **premium high-CTR YouTube-thumbnail visual hierarchy**

while using:

- **Photo 1 exclusively for the left person's facial identity**
- **Photo 2 exclusively for the right person's facial identity**
- **the reference thumbnail—not the uploaded identity photos—to control both people's pose and body language**
- **the user's exact supplied title as the ONLY readable text in the thumbnail**

### FINAL POSE CHECK BEFORE GENERATION

Before rendering, verify:

1. Is the left person's facial identity derived only from Photo 1?
2. Is the right person's facial identity derived only from Photo 2?
3. Has the original pose from Photo 1 been ignored?
4. Has the original pose from Photo 2 been ignored?
5. Do both people follow the conversational posing style of the reference thumbnail?
6. Are their head, shoulder, torso, gaze, and expression angles slightly different from each other?
7. Does the right-side person's pose look noticeably changed from their uploaded input photo?
8. Do the two participants look like they are interacting naturally in the same podcast recording?

If any answer is **no**, correct the pose before generating the final image.

Never generate the final thumbnail until **both identity photos and the exact title have been received**.