Yapper / Talking head

Lik gleda pravo u objektiv i govori brzo, gusto, bez pauze. Nema seta, nema b-rolla, nema montaže osim jump rezova. Ceo format počiva na tome što gledalac oseti da mu se neko obraća direktno i da nema vremena za uvod. Prvi kadar mora da ima pokret, jer statično otvaranje ubija hook rate pre nego što ijedna reč padne.

Write concept, script and explanations in the user's language. **Write all generation prompts in English**, because the models perform better and the user pastes them unchanged.

What it earns, honestly

Prvi frejm odlučuje. Statično otvaranje, logo ili crni title card obara hook rate pre prve reči. Otvaranje sa pokretom subjekta u prvom frejmu nadmašuje statično za oko jedanaest poena hook rate-a. Lik mora već da se kreće kad kadar počne.

Why it persuades

• Direktan pogled aktivira socijalnu pažnju. Lice koje te gleda teže se skroluje od bilo kog objekta.
• Brzina signalizira da informacija ima vrednost. Nema vremena za uvod znači da je gusto.
• Nema produkcije znači nema odbrane. Izgleda kao neko ko ti govori, ne kao neko ko ti prodaje.
• Jump rez je ritam. Svaki rez resetuje pažnju i zabranjuje dosadu.

Stage 1 — Interrogate the product

Ask what the user is selling, at what price, to whom, and what the single claim is. Do not proceed on assumptions. If they already gave enough context, answer for them and show it for confirmation rather than asking again.

The governing rule: the concept is derived FROM the product, never applied TO it. Quality test before delivering: would this same concept work with a competitor's product? If yes, rewrite.

Stage 2 — Pick the style variant

Yapper prodaje bliskošću i direktnim obraćanjem. Crtani ili glineni lik pravi distancu i gledalac prestaje da doživljava obraćanje kao lično. Ako hoćeš animirano direktno obraćanje, to je objašnjivač, ne yapper.

If the user does not choose, recommend one and say why in one line.

Stage 3 — Script the beats

Tri poente, ne četiri. Četvrta se ne pamti i obara hold rate.

Write the actual lines verbatim in the user's language. Do not describe what is said.

Stage 4 — Five hooks against one body

Always deliver five opening variants. When hold rate is fine and hook rate is low, only the first three seconds get regenerated, which is five ads for the price of one production.

Structures that hold for this format:

• „Ako radiš X, gubiš novac i ne znaš.“
• „Tri stvari. Prva ti neće prijati.“
• „Ovo sam naučio pošto sam izgubio [cifra].“
• „Stani. Ovo traje dvanaest sekundi.“
• „Pogrešno te uče da radiš ovo.“
• Počne usred reči, kao da si upao u razgovor

The hook must be a single visible thing. "Curiosity" is not a hook.

Stage 5 — Production

Order of work, do not deviate

1. Generiši osnovni kadar Lice, prostor i svetlo odjednom. Ovo je referenca za sve ostalo.
2. Zaključaj lik Reference Element odmah. Bez toga se lice menja u drugom klipu.
3. Napiši pet hookova pre generacije videa Isti ostatak, pet otvaranja. Generišeš pet puta samo prve tri sekunde.
4. Video kroz Omni Flash Za srpski. Za engleski Veo 3.1.
5. Titluj agresivno Reč po reč, veliki, preko sredine ili u dnu.

Model routing

Shot	Image	Video
Osnovni kadar	Nano Banana Pro	Google Omni Flash za srpski
Isti kadar, engleski	Nano Banana Pro	Veo 3.1
Osoba u pokretu	Nano Banana Pro	Kling 3.0, motion control
Proizvod govori	Nano Banana Pro	Omni Flash ili Veo 3.1

On-camera Serbian goes through Google Omni Flash. Other models carry a foreign accent local audiences detect within half a second. If pronunciation still fails, generate the voice separately and apply lip sync with an audio-driven tool over the finished image rather than accepting a foreign-sounding take.

Locked parameters

Every prompt carries these, explicitly or by clear implication. Changing one means generating a different format.

Prompt skeletons

OSNOVNI KADAR (Nano Banana Pro)

[PERSON DESCRIPTION] framed from the chest up, head near the top edge of frame, looking directly into the lens with eyes locked to the lens rather than past it, already mid-gesture with one hand raised. Behind them an ordinary [ROOM] thrown out of focus, close but unimportant, nothing readable. One soft light source from a window at camera left leaving a gentle shadow on the far cheek. The viewpoint sits at eye level and is static, as if resting on a small tripod. Natural skin texture with visible pores, neutral grade, no cinematic look.

**VIDEO** (Omni Flash za srpski)

[FIRST FRAME] speaks. In [LANGUAGE], fast and dense with no introduction, one small natural verbal stumble: "[LINE]". The head moves naturally with the speech, hands entering the bottom of frame to count. Eyes stay locked to the lens throughout. The camera does not move. Audio: close voice, ordinary room acoustics, no music. A single uninterrupted take filmed in one breath.

**VARIJANTA B** (osoba u pokretu)

[PERSON DESCRIPTION] walking toward the viewpoint along an ordinary [STREET OR CORRIDOR], framed from the chest up, looking directly into the lens with eyes locked to the lens, already speaking as the frame begins. The background is close but unimportant, compressed and thrown out of focus, nothing readable. Natural [TIME OF DAY] light only. The viewpoint moves backward at walking pace, handheld with faint natural shake. Natural skin texture with visible pores, neutral grade, no cinematic look.

### Two phrases that are not optional

Every video prompt closes with `A single uninterrupted take filmed in one breath`, or the models
insert hidden cuts. Never write `zoom`; write `the camera travels toward`. Never name a phone or
camera, or the model draws one into the shot.

---

## Worked example — Proizvod: kurs o AI reklamama. Poenta: agencije naplaćuju produkciju koja više nije potrebna.

**PRIMER · KADAR** (Nano Banana Pro)

A man in his early thirties, dark crew-neck t-shirt, short beard, framed from the chest up, head near the top edge of frame, looking directly into the lens with eyes locked to the lens rather than past it, already mid-gesture with one hand raised. Behind him an ordinary home office thrown out of focus, a shelf and a warm lamp close but unimportant. One soft light source from a window at camera left leaving a gentle shadow on the far cheek. The viewpoint sits at eye level and is static, as if resting on a small tripod. Natural skin texture with visible pores, neutral grade, no cinematic look.

**PRIMER · VIDEO** (Omni Flash)

The frame speaks. In Serbian, fast and dense with no introduction, with one small natural stumble: "Ako još uvek plaćaš nekome da ti snima reklame, ovo... ovo moraš da vidiš." The head moves naturally with the speech, one hand entering the bottom of frame to count. Eyes stay locked to the lens throughout. The camera does not move. Audio: close voice, ordinary room acoustics, no music. A single uninterrupted take filmed in one breath.

Use this to calibrate how much detail belongs in each bracketed slot.

---

## Failures and repairs

**Pogled je pored objektiva**
→ Eksplicitno `eyes locked to the lens, not past it`. Pogled pored kamere ubija format.

**Statičan prvi frejm**
→ Lik mora već da se kreće kad kadar počne. Dodaj `already mid-gesture`.

**Lice se menja kroz klipove**
→ Reference Element lock, i ne menjaj opis svetla između generacija.

**Koža izgleda plastično**
→ Traži `natural skin texture with visible pores` i izbegavaj reč cinematic.

**Prazan zid iza**
→ Uvek obična soba van fokusa. Prazan zid uvek ispadne kao studio.

**Model nacrta telefon u kadru**
→ Nikad ne imenuj telefon ni kameru. Piši `the viewpoint sits`.


**Three-attempt rule.** If the same problem survives two generations, change the approach rather
than the prompt: add a reference image, halve the shot, or split the action into two generations
joined in the edit.

## Assembly

- **Redosled** Hook, kontekst, tri poente sa jump rezovima, dokaz, poziv. Osamnaest sekundi.
- **Jump rezovi** Između svake poente. Rez usred pokreta, ne na pauzi.
- **Titlovi** Veliki, reč po reč, agresivni. Ovo je format koji se najviše gleda bez zvuka.
- **Zvuk** Samo glas i soba. Muzika obara autentičnost.
- **Pet hookova** Generiši pet puta samo prve tri sekunde, ostatak ostaje isti.

---

## Deliver

KONCEPT:      [one paragraph, what happens and why it comes from this product]
VARIJANTA:    [which style, and why]
SCENARIO:     [beat by beat, with verbatim lines]
PET HOOKOVA:  [five openings against the same body]

Then every prompt in English, complete and pasteable, labelled by shot. Then one line on what to
check first in the result.

## Never

- Deliver prompts without a concept behind them.
- Deliver a concept that would survive a product swap.
- Return prompt fragments or advice about what the user should edit. Full prompts only.
- Recommend a style variant this format does not support.
- Let the word "zoom" or a named recording device into a prompt.
- Substitute a model when Serbian speech is on camera.