APIMODELSAPIMODELS

Seedance 2.5 Dialogue & Lip Sync Prompts

Seedance 2.5 video prompts, each shown next to the clip it produced and credited to its author. Where the creator published their prompt we keep it verbatim; where they did not, we reverse-engineer one from the finished clip and label it as such. Hover to play, one click to copy.

← All 95 Seedance 2.5 prompts

Last updated: · Seedance 2.5 API — pricing & playground · API docs

TL;DR

Seedance 2.5 is ByteDance's video model, unveiled on 23 June 2026 at the Volcano Engine FORCE conference. On the shipped API it generates 4 to 30 seconds in a single pass — or duration -1 to let the model choose the length — and accepts up to 50 reference assets in one job: up to 30 images at 4K each, up to 10 video clips and up to 10 audio clips totalling 30 seconds each. It adds video editing and video extension as task types, and generates audio in the same latent space as the picture. Resolution is 480p or 720p; the native 4K at 10-bit shown at the FORCE launch is not exposed on the API. ByteDance also published an official prompting guide for it. This page collects Seedance 2.5 video prompts with the clip each one produced, free to copy. Since 8 August 2026 you can run Seedance 2.5 on apimodels.app as the model seedance-2.5, from $0.120/s; the creators in this library ran these prompts in Dreamina (Jimeng) and Higgsfield.

Source: ByteDance Seed (the model vendor’s official site)

Hover to preview · click for sound

Seven-shot bedroom candid — cat, phone call, doorbell

Montage, multi-shot candid observational footage. Do not use a single camera angle or continuous take. Handheld documentary style with the feeling of accidental real-life capture. Slightly imperfect framing, subtle handheld shake, tiny reframing adjustments, gentle exposure breathing, and autofocus that settles half a beat late. Realistic skin texture, soft indoor natural light, film grain, shallow depth of field. Relaxed breathing, natural blinking, restrained and authentic performance. Total of 7 shots. The woman from @ Image 1 is wearing a matching cotton pajama set consisting of a scoop-neck sleeveless top and loose shorts made from the same fabric and design. Barefoot, she lies on her stomach across a lived-in, slightly messy bed in her bedroom, browsing a mobile shopping page on a black smartphone. A house cat jumps onto the bed seeking attention, and she affectionately plays with it. Moments later, a friend calls. She answers, chats, and eventually bursts into laughter. The doorbell rings, so she ends the call and gets up to collect her dinner. Shot 1 (0–2s): She is already moving. One thumb scrolls through a shopping page while she frowns slightly, thinking. She quietly mutters in Korean: "아, 뭐 살랬더라…" ("Ah... what was I going to buy again?"). Lips synchronize naturally. Her bare feet gently sway behind her. A low handheld camera glides along the edge of the mattress, with soft bedding partially obscuring the foreground. Shot 2 (2–4s): The house cat jumps onto the bed, compressing the blanket as it walks toward her forearm and lets out a single meow. The camera gives a slight jolt from the impact, then quickly reframes from the phone to an over-the-shoulder view that includes her face, hands, phone, and cat. Shot 3 (4–6s): She turns warmly toward the cat, stroking it once from its forehead to its shoulders. In a gentle, affectionate tone she says: "우리 애기 왔어?" ("Did my baby come?"). Lip sync is clear and natural. The camera pushes in briefly through the cat's softly blurred foreground toward her face and hand. Shot 4 (6–8s): A friend's ringtone sounds. She glances down, confirms the caller, swipes to answer, and brings the phone to her ear. The camera moves in a loose semicircle from a slightly tilted overhead angle, capturing the entire answering motion. Shot 5 (8–11s): Her friend begins talking. She responds first with confusion, then disbelief: "어, 왜? 진짜 거짓말." ("Huh? Why? No way, you're kidding."). Immediately afterward she naturally bursts into laughter. The cat kneads the blanket beside her. A close handheld side angle keeps both her phone and profile within the same focal plane. Shot 6 (11–13s): A clear doorbell interrupts the laughter. Both she and the cat turn toward the bedroom door. She braces one hand on the mattress to get up. The camera reacts a fraction too late, briefly turning toward the door before naturally correcting back to her rising movement. Shot 7 (13–15s): Still on the phone, she hurriedly says: "어, 나 밥 왔다. 끊어." ("My food's here. I'll hang up."). She ends the call with her thumb, gets off the bed, and walks toward the door while the cat follows across the blankets. The camera slides low beneath her elbow and tilts upward as she stands in one continuous motion. Sound: No music. Only raw production sound: bedding rustling, quiet breathing, finger taps on the phone, the soft impact of the cat jumping onto the bed, one natural meow, faint purring, a single short incoming ringtone, call connection tone, the friend's voice through the phone, the woman's four Korean lines with accurate lip sync, a genuine brief laugh, one clear doorbell, the tap ending the call, and the soft cat footsteps. No subtitles, no on-screen text, no logo, no watermark. Never render a reference sheet or duplicate the subject.

by oggii_0Source
Hover to preview · click for sound

Two friends talking, Japanese slice of life

A cinematic Japanese slice-of-life scene.Two young Japanese women (age 20 and 22) are sitting together and chatting happily. They are close friends having a relaxed conversation. Their lips move naturally with synchronized dialogue (accurate lip sync).Scene: a quiet residential street in Japan in the late afternoon. Warm golden sunlight, gentle breeze, peaceful everyday

by tanabe_fragmSource
Hover to preview · click for sound

Podcast interview driven by voice references

El entrevistador esta en su podcast entrevistando a @8993a3ce-444a-414b-9800-6798a9c997eb y le pregunta con su voz @0949c9e6-24ae-4895-b567-94aa844e0150: " Buenas mi gente, tenemos un nuevo modelo de IA y hemos traido a un experto en ello, Nabil cuentanos" y entonces el invitado dice con su voz @8993a3ce-444a-414b-9800-6798a9c997eb: "Buenas jordi, la verdad que es una cosa de locos. Gracias a este modelo puedo hacer cosas como esta" y se levanta chasquea los dedos y le sale pelo. Se sienta y dice con su voz @7ac71528-ba91-42ec-92e1-89875ed6e49f: "inténtalo tú" y jordi se levanta y chasquea los dedos y entra un equipo de rugby en el podcast y nabil y jordi empiezan a empujarse con el equipo entero como en una pelea. De repente la camara enfoca a nabil en medio de la pelea y este con su voz @7ac71528-ba91-42ec-92e1-89875ed6e49fdice :" Esto solo es posible con el nuevo modelo de seedance, que loco!" y le da un empujon al que tiene encima forcejeando. De repente jordi en medio de la pelea grita con su voz @0949c9e6-24ae-4895-b567-94aa844e0150: "STOP" y todos los de rugby se convierten en plumas que caen al suelo. Jordi y Nabil se miran y les entra un ataque de risa #seedance2.5 #dreaminacpp @dreamina_ai

by NabilidosooSource
Hover to preview · click for sound

A short film with synchronised dialogue and original score

Create a 30 second cinematic short film titled The Suit Was Still Listening, presented in 16:9 with synchronized dialogue, environmental sound, and original music. The film must feel completely live action, physically grounded, and captured on location, never glossy or synthetic. Visual language: a rain soaked neighborhood laundromat at midnight, warm amber fluorescent tubes inside, cold blue street light and moving car reflections outside, wet pavement, fogged windows, chipped enamel machines, baskets, detergent boxes, loose receipts, realistic skin texture, natural eye moisture, subtle fabric lint, imperfect hair, restrained film grain, gentle halation, lifted shadows, rich but believable color. Wardrobe creates bold editorial color blocking: Mara wears a cobalt trench over a cream blouse, Noah wears a rust knit polo, the teenage attendant wears faded green workwear, and an elderly customer wears a pale gray suit and dark hat. Use expressive medium close shots, tactile inserts, occasional low angles, clean profile compositions, and brief handheld movement that feels operated by a human. Preserve consistent faces, clothing, props, screen direction, reflections, hand anatomy, and spatial continuity. Use realistic speech timing, breathing, blinking, eye focus, and precise lip synchronization. 0 to 3 seconds. Exterior wide shot through falling rain. A lonely laundromat glows on a dark corner. A black sedan turns slowly into the street. Camera makes a subtle forward creep. Music begins with muted upright bass, brushed snare, a low analog pulse, and the rhythmic churn of washers. 3 to 6 seconds. Interior macro shot through a round washer door. A red scarf turns through soap and water. Cut to Mara opening a dryer. A tiny black transmitter falls from the lining of a cream suit and strikes the tile with a sharp metallic click. The score briefly drops out. 6 to 10 seconds. Low medium two shot. Noah freezes beside a folding table. He says quietly, You said clean the suit. Mara studies the transmitter in her palm and replies, I did. You hid a heartbeat in the lining. Keep the delivery controlled and intimate, not theatrical. 10 to 14 seconds. Tight insert on the teenage attendant checking an old signal meter beside the change machine. A green needle jumps. He looks toward the fogged front window and says, It is still transmitting. Camera racks focus from the meter to white headlights sliding across the glass. 14 to 18 seconds. The sedan stops outside. The elderly customer continues folding a shirt without looking up. He says, Then he is already here. Hold on Mara as the moving headlights carve across her face. Let silence, rain, and machine motors carry the tension. 18 to 22 seconds. Noah whispers, Back door. The old man calmly turns the deadbolt on the back exit and answers, Too obvious. Mara notices a delivery rider outside pushing a rolling laundry cart past the window. Use a fast sequence of close shots: her eyes, the transmitter, the cart wheel, the sedan mirror. 22 to 27 seconds. Mara crosses naturally behind a row of machines, slips the transmitter into a sealed laundry bag on the moving cart through an open service hatch, and returns without drawing attention. Noah asks, What did you do. Mara watches the cart continue down the wet street and says, Gave him a cleaner story. 27 to 30 seconds. Exterior telephoto view. The black sedan pulls away and follows the cart. Cut back inside to a symmetrical medium shot of the four characters framed by spinning washer doors. The red scarf circles behind Mara like a slow warning light. She sits opposite Noah and says, Now tell me who owns the suit. End on the transmitter signal fading from the attendant meter as the final bass note lands. Sound design must include rain on glass, distant tires on wet road, washer motors, dryer buttons, fabric movement, the transmitter click, door lock, fluorescent hum, and soft room reflections around every voice. Dialogue must remain clean and naturally mixed above the music. Avoid visual distortion, artificial camera acceleration, floating objects, excessive shallow focus, oversharpening, plastic skin, extra fingers, unreadable signage, random background motion, or dreamlike effects.

by bmx_ai13Source
Hover to preview · click for sound

Alien commander gives orders, in Turkish

Image1 komutan ve lider olan bir uzaylıdır. Image2 ise asker uzaylıdır. komutan askere bir savaş alanında emir veriyor. Karşılıklı diyalog sahnesi. Konuşmalar tamamen uydurma bir uzaylı dilinde olmalı Bir blockbuster filminden bir savaş sahnesi olmalı. Sinematik kamera hareketleri ve açıları kullanmalısın.

by ozansihaySource

How to write great Seedance 2.5 prompts

  1. 1Write it in one fixed order. Subject, then action, then location, then camera work, then lighting, then picture texture. "A person walking" gets you a stock loop; the same subject with a camera move and a light direction attached gets you a shot. Creators who have spent real credits on 2.5 converge on this order.
  2. 2Thirty seconds is written as stages, not as longer sentences. The official guide splits a long clip into [STAGE ONE] / [STAGE TWO] / [STAGE THREE], each with an opening state, one main event, and an end state the next stage inherits. Padding a 10-second prompt with adjectives does not make a 30-second clip; splitting the event into inherited states does.
  3. 3One main action per time segment, described exactly once. Whether you use stages or plain 0-5s / 5-10s ranges, give each segment a single main movement. Two things go wrong otherwise: three actions crammed into five seconds all get compressed, and the guide is explicit that describing the same action twice degrades it.
  4. 4Give every reference asset a job and an exclusion. Not "here are my images" but "@image1 supplies the armour design and colour breakdown. Do not take the background." You can attach up to 50 assets; the more you attach without assigning roles, the faster stability drops.
  5. 5Name your subjects and reuse the name. Write <SOLDIER>, <SHARI>, <WOMAN> once and refer to that token every time after. Over 30 seconds, "she" and "the man" drift onto the wrong body — a named token is the most affordable identity lock there is.
  6. 6Write the sound as a cue sheet, not as a mood. Audio comes out of the same pass as the picture, so list what you want: ambience, the specific foley, dialogue with the language named, and what should be absent. Several prompts in this library end with a full sound list, and it is a large part of why they look finished.
  7. 7Ask for the imperfections you want. Handheld shake, autofocus that settles late, exposure breathing, framing that clips a face. Realism in this model is a list of specified flaws — leave them out and you get a clean commercial no matter how many times you write "candid".
  8. 8End with the negative list, and keep parameters out of it. Close with what must not appear: no subtitles, no watermark, no extra limbs, and — for reference-driven jobs — never render the reference sheet itself or duplicate the subject. Resolution, duration and aspect ratio are generation settings, not prompt text.

Prompt formula

Subject + action + location + camera work + lighting + picture texture + sound. Past about ten seconds, wrap that in timed stages: each stage gets one main action and an end state the next stage inherits, then close with the negative list.

Example: <BARISTA> pulls the last shot of the night in a closing café and sets the cup on the pass. Rain on the window, chairs already stacked, one warm pendant lamp over the machine and the rest of the room in shadow. The camera holds a low medium on the portafilter, then rises to a profile as the cup lands. Photoreal, warm tungsten against cold window light, fine grain. Sound: the grinder winding down, steam hiss, rain, one mug on wood, no music. Do not add customers, subtitles or a watermark.

Frequently asked questions

What is Seedance 2.5?

Seedance 2.5 is ByteDance's video generation model, unveiled on 23 June 2026 at the Volcano Engine FORCE conference. On the shipped API it generates 4 to 30 seconds in a single pass (or duration -1 to let the model pick the length), takes up to 50 multimodal reference assets in one job (up to 30 images, up to 10 video clips and 10 audio clips), adds video editing and video extension as task types so a product or background can be swapped without re-rolling the whole clip, and generates audio in the same latent space as the picture. Output is 480p or 720p — the native 4K at 10-bit colour announced at FORCE is not something the API exposes. ByteDance also shipped an official prompting guide for it, which we have walked through in English.

Are these prompts free to use?

Yes. Every prompt on this page is free to copy and run anywhere you have access to Seedance 2.5 — no login, no payment.

Who wrote these prompts?

Two kinds, and they are labelled differently. Prompts the creator published are kept in the language they wrote them in, credited, and linked to the original post — copyright stays with the author. For clips whose creator never published a prompt, we reverse-engineer one from the finished video and mark it "reconstructed": it describes the output, so it cannot recover negative constraints, exact dialogue or the reference-image workflow. It is a writing reference, not the creator's prompt. If you are an author and want an entry removed or a credit corrected, email us.

Where can I run Seedance 2.5 today?

On apimodels.app, as the model seedance-2.5, since 8 August 2026 — 480p $0.120/s, 720p $0.270/s, charged only on success. That is a change: probing our production Volcengine Ark account on 4 August 2026, every Seedance 2.5 model id still returned InvalidEndpointOrModel.NotFound, and we said so rather than sell a page that errors on generate. One key covers it alongside Seedance 2.0 (still the one to use if you need 1080p, from $0.092/s), MiniMax H3, Kling, VEO and Grok. The creators in this library ran these prompts in ByteDance's own Dreamina (Jimeng) and in Higgsfield.

How is writing for Seedance 2.5 different from Seedance 2.0?

The craft is the same; the structure is not. At 4-15 seconds you can write one paragraph for one action. At 30 seconds you have to split the clip into stages, give each stage one main action and an end state the next stage inherits, and lock identity with named subject tokens instead of pronouns — otherwise the model spreads the action evenly and identity drifts across the back half. The reference workflow also scales: 2.0 takes up to 9 images plus a reference video and audio, 2.5 takes up to 50 assets, which makes assigning each one an explicit job and exclusion much more important.

How long should a Seedance 2.5 prompt be?

As long as the clip needs and no longer. The prompts in this library run from about 1,700 to 4,600 characters, and the longest ones are the 30-second pieces — not because they are wordier, but because a 30-second clip has more states to pin down. If your prompt is long because it describes the same action three ways, cut it; the official guide is explicit that repeating an action degrades it.

Run these prompts on apimodels.app

Seedance 2.5 is live here alongside 160+ other image, video, audio and language models — one API, one key, pay as you go, with $0.10 of trial credit on sign-up.

This library is open source on GitHub

The prompts we reconstructed are published in full under MIT; author-written prompts are indexed with credit and a link to the original post. Open an issue to suggest more.

View on GitHub

Other prompt libraries