Section 01
THE SLATE | How The Spokesperson Works
The Spokesperson, Vol.M
What this volume is
Six vertical thirty-second ads, one per Vol.22 brand, fronted by one presenter who talks to camera for the whole thirty seconds with the product in her hand. This is the format brands pay creators for right now, and it is the hardest thing the model can do: a face that holds identity through a full script, lips in sync for thirty seconds, a hand that never breaks, a product that never drifts, in 9:16. Vol.L gave the character one line. Vol.M gives her the whole pitch.
Each ad is two files. A talking take, thirty seconds, where she delivers the script. A cut-in, ten seconds, product only, that you drop over the demo beat in the edit so the ad breathes and the viewer sees the product close. Captions and the endcard are real type added in post, never generated. Six ads, six scripts, twelve prompts, every one on the page.
The six rules this volume adds
1. The Presenter Bible. One face made once with text-to-image, one body made from that face with image-to-image, both on plain grey, both in a plain white tee so no ad’s wardrobe fights the product. The same two images carry all six ads. A presenter is a brand asset; you build her once.
2. The Script Grid. Thirty seconds is five beats, timed: Hook 0 to 4, Problem 4 to 9, Demo 9 to 19, Proof 19 to 25, Ask 25 to 30. Under seventy words in total, in braces, one sentence per beat, written the way a person talks. The model can hold thirty seconds of speech if the words are few and the beats are marked; it drops sync when the script is a paragraph.
3. The Hand Rule. The product is in her hand from the first frame to the last, label to camera, and every beat says what the hand does: holds, turns, opens, lifts, sets down. Exactly two hands, five fingers each. A hand that has nothing to do is a hand the model gets wrong.
4. The Vertical Frame. 9:16 with the presenter’s eyes on the upper third line, the product at chest height in the lower half, and the top fifth and bottom fifth kept quiet for captions and the platform’s own overlays. Write the frame, not just the shot.
5. The Cut-in. A separate ten-second product-only take with no person, same product reference, same palette, shot close. It covers the demo beat in the edit, hides any sync wobble, and gives every ad its macro. Cheap, and it makes the ad look produced.
6. The Caption Plate. No text is generated. Captions are burned in post from the script, one beat at a time, in the brand’s type. The endcard is the final frame plus the wordmark and the ask in real type. The Clean Plate Rule from Vol.J, applied to a talking film.
Carried over unchanged
- Reference binding with ONLY and Do NOT, one line per image, the count line for the person and the product.
- The Brand Lock: the wordmark reads exactly what the Vol.22 image says, every frame.
- One camera behaviour per take. Here it is fixed: a locked handheld phone frame, no moves, the energy is in her.
- End states on every beat. Settle and hold on the last one.
- 720p, never 1080p. Draft at 480p only if a take misbehaves, then lock and render.
The six ads
- AD01 SOLA, the cold brew: the morning-person hook. Teaches the Script Grid.
- AD02 EMBER, the coffee bag: the "I switched" hook. Teaches the Hand Rule with a bag that opens on camera.
- AD03 KITE, the running shoe: the "stop scrolling" hook. Teaches the Vertical Frame with a product that lives at the bottom of the frame.
- AD04 NOOR, the serum: the "two minutes" routine hook. Teaches the Cut-in with a dropper macro.
- AD05 VESPER, the fragrance: the gift hook. Teaches speaking softly, low light, and a product that must not glow.
- AD06 OATH, the granola: the "what I eat before" hook. Teaches the Ask and the Caption Plate.
Thirty-second anatomy
Cost per ad at house settings
One thirty-second take at 720p in 9:16 is about 3,900 credits and the ten-second cut-in about 1,300, so one ad is roughly 5,200 and the six about 31,000, plus about 70 for the bible. Draft the structure at 480p for a third of that if a script fights you, then render the locked prompt at 720p.
Delivery, every ad
Section 02
AD01 | SOLA, The Morning Person
The Spokesperson, Vol.M
The first ad is the template. Five beats, sixty-two words, a bottle that never leaves her hand.
AD01 teaches the Script Grid, the rule every other ad in the volume stands on. A thirty-second talking take works when the words are few and every beat is timed and marked; it collapses into a mumble when you hand the model a paragraph. So the script is five lines, one per beat, in braces, under seventy words, and each beat also says what her hand does and where the bottle ends up. The model is animating a mouth and a hand at the same time; you have to give it both.
The script grid
The presenter bible
Made once for the whole volume: face with text-to-image, body from that face with image-to-image, plain white tee, plain grey. Both prompts are on the Slate. Use the same two images as @Image 1 and @Image 2 in all six talking takes.
Talking take, 30 seconds
Cut-in, 10 seconds
The edit
Lay the talking take down in full. Over seconds 11 to 17 of the demo beat, place the cut-in as a picture-only insert, her voice continuing underneath. Burn the captions from the script, one beat at a time, in the brand’s grotesque sans, in the quiet bottom fifth. Endcard: the final frame, the SOLA wordmark and "Link in bio" in real type, three seconds.
What to check on the still frames
- At 4s: bottle raised, label to camera, one person, eyes on the upper third.
- At 12s: the cap off, the label filling the lower centre.
- At 22s: the bottle still and close, SOLA readable.
- At 30s: her mouth closed, the bottle at chest height, nothing moving.
Fix Bank
- Lips drift from the words after fifteen seconds: the script is over seventy words or a beat has two sentences; cut words before cutting anything else.
- The bottle leaves her hand: "in her hand from the first frame to the last" is missing from the product line; the Hand Rule is a binding, not a stage direction.
- Six fingers: the count line is missing from FRAME.
- The camera drifts: "one locked handheld phone frame, no camera moves" must be in the first sentence and "camera moves" in DO NOT.
- Music appears: "no music" in AUDIO and "music" in DO NOT.
- She says words that are not in the script: "extra words beyond the braces" in DO NOT; keep the VOICE line, "the lines in braces are the only speech".
- The label re-spells in the demo beat: the bottle is too far from the lens; "lifts the bottle toward the lens so the label fills the lower centre" is the fix.
- The kitchen is the Vol.22 kitchen: the ban on the marble counter and tiled wall is missing from the @Image 3 line.
Section 03
AD02 | EMBER, I Switched
The Spokesperson, Vol.M
A bag that opens on camera. This ad teaches the Hand Rule.
The SOLA bottle in AD01 was easy for the hands: one object, one grip. A stand-up bag is the real test. It folds, it has a tie, it opens, and a model that is not told exactly what the hands do will give you a bag that morphs, a tie that vanishes, or a third hand. The Hand Rule is that every beat names the action of the hand and the end state of the product, and the product line binds it to her hands for the whole take. Watch the demo beat: unfold the tie, open the top, lift toward the lens. Three actions, in order, with the wordmark kept visible through all of them.
The script grid
Talking take, 30 seconds
Cut-in, 10 seconds
The edit
Talking take in full; the cut-in over seconds 12 to 18 while she says "Open it"; captions from the script, one beat at a time, in the brand’s bold grotesque, in the bottom fifth. Endcard: the final frame, the EMBER wordmark and "Roasted Tuesday" in real type, three seconds.
What to check on the still frames
- At 4s: the bag in both hands, EMBER to camera, one person.
- At 14s: the tie unfolded, the bag open, the wordmark still readable.
- At 22s: the bag closed and still, EMBER sharp.
- At 30s: her mouth closed, the bag at chest height, nothing moving.
Fix Bank
- The bag changes shape when it opens: add "the bag keeps its shape, only the top opens" to the demo beat.
- The tin tie disappears: name it in every beat where it is handled, and keep "the copper tin tie" in the product line.
- A third hand appears during the open: the count line "exactly two hands, five fingers each" must be in FRAME; add "her left hand holds the bag, her right hand works the tie".
- The wordmark hides behind her fingers: "the wordmark still visible" in the demo beat; if it persists, "she holds the bag by its lower corners".
- She smells the bag with her eyes closed for too long: add "eyes back to lens by the end of the beat" as the end state.
- Beans spill on the counter: "no beans on the counter" in FRAME; the ban on the Vol.22 beans is in the product line.
- Lips drift on the demo beat: it has three sentences, the most in the volume; cut "honestly" first if it wobbles.
Section 04
AD03 | KITE, Stop Scrolling
The Spokesperson, Vol.M
A product that wants to live at the bottom of the frame. This ad teaches the Vertical Frame.
A shoe is the wrong shape for a phone screen: wide, low, and it wants to sit at her feet where the platform’s own buttons live. The Vertical Frame rule puts the presenter’s eyes on the upper third line, the product at chest height in the lower half, and keeps the top fifth and the bottom fifth quiet for captions and overlays. So the shoe is held up, side-on, never on the floor, and the frame is written before the shot: where the eyes are, where the product is, what stays empty.
The script grid
Talking take, 30 seconds
Cut-in, 10 seconds
The edit
Talking take in full; the cut-in over seconds 11 to 17 while she turns the shoe; captions from the script in the brand’s bold grotesque, bottom fifth. Endcard: the final frame, the KITE mark and "Run quiet" in real type, three seconds.
What to check on the still frames
- At 4s: the shoe side-on at chest height, coral panel and KITE to camera, no swoosh.
- At 14s: the sole shown, then side-on again, her thumb on the panel.
- At 22s: the shoe still and close, KITE readable.
- At 30s: her mouth closed, the shoe at chest height, nothing moving.
Fix Bank
- The shoe drops to the floor: the Hand Rule binding "in her hand from the first frame to the last" is missing; add "never on the floor".
- A swoosh appears: ban it by shape, "any curved stripe or swoosh shape", never by brand name.
- She is wearing the shoes: "Do NOT reproduce the shoes in @Image 2" plus "she holds exactly this one shoe".
- The bottom fifth fills with floor and her feet: "the top fifth and bottom fifth quiet" in FRAME; raise the shoe to chest height in every beat.
- The shoe turns into two shoes on the turn: "a second shoe" in DO NOT and "exactly this one shoe" in the product line.
- Her hair covers the earrings and the face drifts: keep the face line first; add "her hair stays off her face".
Section 05
AD04 | NOOR, Two Minutes
The Spokesperson, Vol.M
The demo touches her face. This ad teaches the Cut-in.
Skincare demos are the riskiest thirty seconds in the volume: a pipette, two drops, fingertips on a cheek, all while she keeps talking and the wordmark stays readable. It will not be perfect every time, and it does not have to be. The Cut-in rule is that every ad carries a separate ten-second product-only take, shot close, that you lay over the demo beat in the edit. Her voice keeps running underneath; the viewer sees the drop form and fall in macro instead of a hand that wobbled. It costs a third of a talking take and it is the difference between a demo and a produced ad.
The script grid
Talking take, 30 seconds
Cut-in, 10 seconds
The edit
Talking take in full; the cut-in over seconds 10 to 16, exactly where the drops fall in her take, so the macro replaces the hardest part of the demo; captions from the script in thin capitals matching the wordmark, bottom fifth. Endcard: the final frame, the NOOR wordmark and "Two minutes. Every morning." in real type, three seconds.
What to check on the still frames
- At 4s: the bottle raised, NOOR to camera, one person.
- At 13s: two drops on her fingertips, the bottle still in her left hand.
- At 22s: the pipette back on, the bottle still and close, NOOR readable.
- At 30s: her mouth closed, the bottle at chest height, nothing moving.
Fix Bank
- The bottle vanishes during the press: "the bottle still in her left hand, wordmark visible" must be inside the demo beat, not only in the binding.
- The drops become a stream: "two drops" in the beat and "exactly two drops" if it persists.
- She rubs both cheeks and the take runs long: "one cheek" and the end state "her fingertips leaving her cheek".
- The pipette becomes a pump: keep "the white pipette cap" in the product line and "unscrews the pipette" in the demo.
- The tiles read as the Vol.22 set: the ban on the linen and the eucalyptus is in the product line; add "white tiles behind her" to FRAME.
- Lips drift on the demo beat: it is the longest line in the volume by design; if it drifts, the cut-in covers it, which is the point of this page.
Section 06
AD05 | VESPER, The Gift
The Spokesperson, Vol.M
Low light, a soft voice, a product that must not glow. This ad teaches restraint.
Everything in the first four ads is bright and quick. A fragrance ad is the opposite, and the model fights you on all three counts: it lifts the exposure until the dusk is noon, it turns a whisper into a presenter voice, and it makes frosted glass glow like a lamp. So the take names its one light (the dresser lamp, on one side of her face), the VOICE line asks for almost a whisper and the AUDIO line bans music, and the product line states that the bottle does not emit light, twice, once in the binding and once in DO NOT. Restraint is written, not hoped for.
The script grid
Talking take, 30 seconds
Cut-in, 10 seconds
The edit
Talking take in full; the cut-in over seconds 12 to 18 while she sprays; captions in the brand’s high-contrast serif, small, bottom fifth. Endcard: the final frame, the VESPER wordmark and "One spray. Then the night." in real type, three seconds.
What to check on the still frames
- At 4s: the bottle raised, the brass lit, one side of her face in lamp light, the dusk still blue.
- At 14s: the cap off, one spray beside her, the bottle still in her left hand.
- At 22s: the bottle still and close, VESPER readable, not glowing.
- At 30s: her mouth closed, the bottle at chest height, nothing moving.
Fix Bank
- The room becomes daylight: name the lamp as the only light in FRAME and keep "blue evening light at the window" in the first sentence.
- The bottle glows: "does not glow or emit light" in the product line and "a glowing bottle" in DO NOT, both.
- She speaks at full volume: the VOICE line must say "almost a whisper, close to the phone".
- The spray hits her face: "sprays once into the air beside her at shoulder height".
- The brass cap turns gold plastic: keep "brushed brass" in the product line and "the brass catching the lamp" in FRAME.
- The take is too dark to read the wordmark: add "the lamp also lights the label" to FRAME; the cut-in covers the rest.
Section 07
AD06 | OATH, What I Eat Before
The Spokesperson, Vol.M
The last ad teaches the Ask and the Caption Plate, the two things that turn a take into an ad.
Five ads in, the mechanics are yours: the bible, the grid, the hand, the frame, the cut-in. AD06 is about the finish. The Ask is five words at twenty-five seconds, the product settling, a nod, a hold; never a speech, never a second call. The Caption Plate is everything you add in post: captions burned from the script one beat at a time in the brand’s type, in the quiet bottom fifth, and an endcard built from the final frame with the wordmark and the ask in real type. No text is ever generated. The model gives you a person and a product; you give it a brand.
The script grid
Talking take, 30 seconds
Cut-in, 10 seconds
The Caption Plate, step by step
- Import the talking take. Drop the cut-in over seconds 11 to 17 as picture only.
- Captions: one text layer per beat, the exact words from the grid, the brand’s bold grotesque sans, white with a soft shadow, in the bottom fifth, on for the length of the beat.
- Endcard: freeze the final frame for three seconds; add the OATH wordmark exactly as it appears on the pouch, and under it "Fuel first." in the same family.
- Export 1080 by 1920 at 30 fps, 33 seconds. Name it VOLM_AD06_Final.mp4.
What to check on the still frames
- At 4s: the pouch in both hands, OATH and the window to camera, one person.
- At 14s: the seal open, a handful taken, the pouch held toward the lens with the wordmark readable.
- At 22s: the pouch closed and still, OATH sharp.
- At 30s: her mouth closed, the pouch at chest height, nothing moving.
Fix Bank
- She talks with her mouth full and the sync breaks: "her mouth free" as the end state of the demo, and keep the handful small.
- The pouch collapses when opened: "the pouch keeps its shape, only the seal opens".
- The window shows nothing: "the golden granola visible inside" in the product line and "the window to camera" in FRAME.
- The lamp becomes daylight: keep "before sunrise, one warm lamp on, blue light at the window" in the first sentence.
- The Ask runs long: five words, one nod, one hold; if she adds a sentence, "extra words beyond the braces" is missing from DO NOT.
- Captions cover her mouth: they belong in the bottom fifth; if the framing put her lower, re-run the take with FRAME first, do not move the captions up.
The volume, closed
Six ads, one presenter, six products, six scripts under seventy words, twelve takes, six endcards. When a brand asks you for a creator ad, this is the recipe: build the presenter once, write the grid, bind the hand, frame vertical, cover the demo, finish in post. Change the product and the script, keep the system.