MODEL IDopenai:3@2
deprecated

Sora 2 Pro

OpenAI
by OpenAI

Sora 2 Pro is the higher quality Sora 2 variant for precision video work. It supports text prompts and image inputs. It outputs synchronized video with sound, higher resolution frames, and stronger temporal consistency. Ideal for production clips and demanding pipelines.

Sora 2 Pro

This model is deprecated and will be deactivated on September 24, 2026.

text-to-video

Seismic Journal Logo Loop

import { createClient } from '@runware/sdk'

const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()

const [result] = await client.run({
  model: 'openai:3@2',
  positivePrompt: 'Create a polished 12-second seamless logo loop for FAULTLINE, an independent earthquake-engineering journal. Full-frame 16:9 motion graphic, not a mockup. The setting is a warm ivory letterpress-paper field viewed directly from above. Centered on the field is a precise editorial logo lockup: the exact uppercase word “FAULTLINE” in bold condensed black sans-serif lettering, paired with a small safety-orange square containing one clean diagonal fault line. No other words, captions, borders, or interface elements.\n\nBegin with the complete logo perfectly still and sharply registered. After one second, a thin graphite-black baseline extends beneath the wordmark and becomes a restrained seismograph trace. The pulse travels left to right, making the individual letters shift apart by a few millimeters like heavy movable-type blocks responding to a tremor. Fine paper fibers lift, tiny grains of graphite jump, and the orange square divides along its diagonal fault before sliding back together. Keep every letter intact, correctly spelled, upright, and readable throughout. The movement should feel engineered and controlled rather than destructive. The waveform gradually loses amplitude, the type blocks settle into exact alignment, and the baseline retracts cleanly. By the final second, restore the opening logo and paper texture exactly so the end joins the first frame invisibly.\n\nLocked overhead camera with no cuts, zooms, tilts, or perspective drift. Tactile stop-motion letterpress aesthetic combined with rigorous Swiss editorial design. Limited palette of warm ivory, dense carbon black, and vivid safety orange. Soft diffuse studio illumination, subtle debossed shadows, crisp edges, premium print texture, generous negative space, restrained confident mood. Synchronized sound design: delicate graphite scratching, muted wooden type clicks, one low controlled bass thump at the main pulse, then quiet room tone; no speech and no music.',
  width: 1920,
  height: 1080,
  duration: 12
})
import asyncio
import os

from runware import Runware


async def main():
    async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
        results = await client.run({
            "model": "openai:3@2",
            "positivePrompt": "Create a polished 12-second seamless logo loop for FAULTLINE, an independent earthquake-engineering journal. Full-frame 16:9 motion graphic, not a mockup. The setting is a warm ivory letterpress-paper field viewed directly from above. Centered on the field is a precise editorial logo lockup: the exact uppercase word “FAULTLINE” in bold condensed black sans-serif lettering, paired with a small safety-orange square containing one clean diagonal fault line. No other words, captions, borders, or interface elements.\n\nBegin with the complete logo perfectly still and sharply registered. After one second, a thin graphite-black baseline extends beneath the wordmark and becomes a restrained seismograph trace. The pulse travels left to right, making the individual letters shift apart by a few millimeters like heavy movable-type blocks responding to a tremor. Fine paper fibers lift, tiny grains of graphite jump, and the orange square divides along its diagonal fault before sliding back together. Keep every letter intact, correctly spelled, upright, and readable throughout. The movement should feel engineered and controlled rather than destructive. The waveform gradually loses amplitude, the type blocks settle into exact alignment, and the baseline retracts cleanly. By the final second, restore the opening logo and paper texture exactly so the end joins the first frame invisibly.\n\nLocked overhead camera with no cuts, zooms, tilts, or perspective drift. Tactile stop-motion letterpress aesthetic combined with rigorous Swiss editorial design. Limited palette of warm ivory, dense carbon black, and vivid safety orange. Soft diffuse studio illumination, subtle debossed shadows, crisp edges, premium print texture, generous negative space, restrained confident mood. Synchronized sound design: delicate graphite scratching, muted wooden type clicks, one low controlled bass thump at the main pulse, then quiet room tone; no speech and no music.",
            "width": 1920,
            "height": 1080,
            "duration": 12
        })


asyncio.run(main())
curl https://api.runware.ai/v1 \
  -H "Authorization: Bearer $RUNWARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '[
    {
      "taskType": "videoInference",
      "taskUUID": "bef00999-94dc-4b66-a812-39c451aea193",
      "model": "openai:3@2",
      "positivePrompt": "Create a polished 12-second seamless logo loop for FAULTLINE, an independent earthquake-engineering journal. Full-frame 16:9 motion graphic, not a mockup. The setting is a warm ivory letterpress-paper field viewed directly from above. Centered on the field is a precise editorial logo lockup: the exact uppercase word “FAULTLINE” in bold condensed black sans-serif lettering, paired with a small safety-orange square containing one clean diagonal fault line. No other words, captions, borders, or interface elements.\n\nBegin with the complete logo perfectly still and sharply registered. After one second, a thin graphite-black baseline extends beneath the wordmark and becomes a restrained seismograph trace. The pulse travels left to right, making the individual letters shift apart by a few millimeters like heavy movable-type blocks responding to a tremor. Fine paper fibers lift, tiny grains of graphite jump, and the orange square divides along its diagonal fault before sliding back together. Keep every letter intact, correctly spelled, upright, and readable throughout. The movement should feel engineered and controlled rather than destructive. The waveform gradually loses amplitude, the type blocks settle into exact alignment, and the baseline retracts cleanly. By the final second, restore the opening logo and paper texture exactly so the end joins the first frame invisibly.\n\nLocked overhead camera with no cuts, zooms, tilts, or perspective drift. Tactile stop-motion letterpress aesthetic combined with rigorous Swiss editorial design. Limited palette of warm ivory, dense carbon black, and vivid safety orange. Soft diffuse studio illumination, subtle debossed shadows, crisp edges, premium print texture, generous negative space, restrained confident mood. Synchronized sound design: delicate graphite scratching, muted wooden type clicks, one low controlled bass thump at the main pulse, then quiet room tone; no speech and no music.",
      "width": 1920,
      "height": 1080,
      "duration": 12
    }
  ]'
runware run openai:3@2 \
  positivePrompt="Create a polished 12-second seamless logo loop for FAULTLINE, an independent earthquake-engineering journal. Full-frame 16:9 motion graphic, not a mockup. The setting is a warm ivory letterpress-paper field viewed directly from above. Centered on the field is a precise editorial logo lockup: the exact uppercase word “FAULTLINE” in bold condensed black sans-serif lettering, paired with a small safety-orange square containing one clean diagonal fault line. No other words, captions, borders, or interface elements.

Begin with the complete logo perfectly still and sharply registered. After one second, a thin graphite-black baseline extends beneath the wordmark and becomes a restrained seismograph trace. The pulse travels left to right, making the individual letters shift apart by a few millimeters like heavy movable-type blocks responding to a tremor. Fine paper fibers lift, tiny grains of graphite jump, and the orange square divides along its diagonal fault before sliding back together. Keep every letter intact, correctly spelled, upright, and readable throughout. The movement should feel engineered and controlled rather than destructive. The waveform gradually loses amplitude, the type blocks settle into exact alignment, and the baseline retracts cleanly. By the final second, restore the opening logo and paper texture exactly so the end joins the first frame invisibly.

Locked overhead camera with no cuts, zooms, tilts, or perspective drift. Tactile stop-motion letterpress aesthetic combined with rigorous Swiss editorial design. Limited palette of warm ivory, dense carbon black, and vivid safety orange. Soft diffuse studio illumination, subtle debossed shadows, crisp edges, premium print texture, generous negative space, restrained confident mood. Synchronized sound design: delicate graphite scratching, muted wooden type clicks, one low controlled bass thump at the main pulse, then quiet room tone; no speech and no music." \
  width=1920 \
  height=1080 \
  duration=12
{
  "taskType": "videoInference",
  "taskUUID": "bef00999-94dc-4b66-a812-39c451aea193",
  "model": "openai:3@2",
  "positivePrompt": "Create a polished 12-second seamless logo loop for FAULTLINE, an independent earthquake-engineering journal. Full-frame 16:9 motion graphic, not a mockup. The setting is a warm ivory letterpress-paper field viewed directly from above. Centered on the field is a precise editorial logo lockup: the exact uppercase word “FAULTLINE” in bold condensed black sans-serif lettering, paired with a small safety-orange square containing one clean diagonal fault line. No other words, captions, borders, or interface elements.\n\nBegin with the complete logo perfectly still and sharply registered. After one second, a thin graphite-black baseline extends beneath the wordmark and becomes a restrained seismograph trace. The pulse travels left to right, making the individual letters shift apart by a few millimeters like heavy movable-type blocks responding to a tremor. Fine paper fibers lift, tiny grains of graphite jump, and the orange square divides along its diagonal fault before sliding back together. Keep every letter intact, correctly spelled, upright, and readable throughout. The movement should feel engineered and controlled rather than destructive. The waveform gradually loses amplitude, the type blocks settle into exact alignment, and the baseline retracts cleanly. By the final second, restore the opening logo and paper texture exactly so the end joins the first frame invisibly.\n\nLocked overhead camera with no cuts, zooms, tilts, or perspective drift. Tactile stop-motion letterpress aesthetic combined with rigorous Swiss editorial design. Limited palette of warm ivory, dense carbon black, and vivid safety orange. Soft diffuse studio illumination, subtle debossed shadows, crisp edges, premium print texture, generous negative space, restrained confident mood. Synchronized sound design: delicate graphite scratching, muted wooden type clicks, one low controlled bass thump at the main pulse, then quiet room tone; no speech and no music.",
  "width": 1920,
  "height": 1080,
  "duration": 12
}
Response
{
  "taskType": "videoInference",
  "taskUUID": "bef00999-94dc-4b66-a812-39c451aea193",
  "videoUUID": "c8f57959-b3c0-40e8-a0a0-c621c4042d1f",
  "videoURL": "https://vm.runware.ai/video/os/a03d21/ws/5/vi/c8f57959-b3c0-40e8-a0a0-c621c4042d1f.mp4",
  "seed": 1820682051,
  "cost": 8.4,
  "outputs": {
    "videoId": "video_6a960c155d588193b58c0638de13fe580f24ea3c171e8bc5"
  }
}
text-to-video

Seed Archive Synthpop Video

import { createClient } from '@runware/sdk'

const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()

const [result] = await client.run({
  model: 'openai:3@2',
  positivePrompt: 'Create a finished 12-second, 16:9 music-video segment for an original art-pop single. Inside a vast, immaculate agricultural seed archive, long stainless-steel sorting tables and analog conveyor belts form strong geometric lines beneath cream-colored fluorescent ceiling panels. Thousands of dry beans and corn kernels in rust, mustard, ivory, black, and dusty blue travel across the machinery. The lead singer is a poised woman in her early thirties wearing a sharply tailored cobalt work jacket, ochre shirt, and dark trousers; four seed technicians in coordinated off-white uniforms work behind her.\n\nOne continuous choreographed shot with no cuts. Begin in a symmetrical wide shot looking straight down the central aisle as the singer walks toward camera. The camera smoothly dollies backward at her pace with subtle 35mm handheld texture. On each beat, the technicians perform crisp, believable sorting gestures—tilting steel trays, sliding archival drawers, and releasing streams of seeds that create graphic waves across the tables. At four seconds, the singer reaches foreground and sings directly into the lens with accurate, natural lip synchronization: “Count every kernel, keep a little summer, sealed beneath the mountain till the green returns.” During “little summer,” the camera makes a gentle half-orbit around her while a fan lifts the corner of her jacket; the moving seed colors create rhythmic bands in the background. End on a medium close-up as she holds eye contact and the conveyors stop together on the final word.\n\nVisual direction: 1970s agricultural educational film reimagined as refined contemporary fashion cinema; clean production design, restrained deadpan choreography, warm fine-grain 16mm texture, slightly faded cream highlights, rich cobalt and ochre accents, realistic skin and materials, soft halation around practical lights. Controlled, confident, quietly hopeful mood. Original synchronized audio: dry analog drum-machine groove at 112 BPM, rubbery mono synth bass, muted handclaps, small metallic conveyor percussion, and an intimate clear female alto singing the specified lyric. Natural room resonance and subtle machine sounds integrated into the track. No captions, titles, logos, labels, watermarks, extra dialogue, or visible writing.',
  width: 1920,
  height: 1080,
  duration: 12
})
import asyncio
import os

from runware import Runware


async def main():
    async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
        results = await client.run({
            "model": "openai:3@2",
            "positivePrompt": "Create a finished 12-second, 16:9 music-video segment for an original art-pop single. Inside a vast, immaculate agricultural seed archive, long stainless-steel sorting tables and analog conveyor belts form strong geometric lines beneath cream-colored fluorescent ceiling panels. Thousands of dry beans and corn kernels in rust, mustard, ivory, black, and dusty blue travel across the machinery. The lead singer is a poised woman in her early thirties wearing a sharply tailored cobalt work jacket, ochre shirt, and dark trousers; four seed technicians in coordinated off-white uniforms work behind her.\n\nOne continuous choreographed shot with no cuts. Begin in a symmetrical wide shot looking straight down the central aisle as the singer walks toward camera. The camera smoothly dollies backward at her pace with subtle 35mm handheld texture. On each beat, the technicians perform crisp, believable sorting gestures—tilting steel trays, sliding archival drawers, and releasing streams of seeds that create graphic waves across the tables. At four seconds, the singer reaches foreground and sings directly into the lens with accurate, natural lip synchronization: “Count every kernel, keep a little summer, sealed beneath the mountain till the green returns.” During “little summer,” the camera makes a gentle half-orbit around her while a fan lifts the corner of her jacket; the moving seed colors create rhythmic bands in the background. End on a medium close-up as she holds eye contact and the conveyors stop together on the final word.\n\nVisual direction: 1970s agricultural educational film reimagined as refined contemporary fashion cinema; clean production design, restrained deadpan choreography, warm fine-grain 16mm texture, slightly faded cream highlights, rich cobalt and ochre accents, realistic skin and materials, soft halation around practical lights. Controlled, confident, quietly hopeful mood. Original synchronized audio: dry analog drum-machine groove at 112 BPM, rubbery mono synth bass, muted handclaps, small metallic conveyor percussion, and an intimate clear female alto singing the specified lyric. Natural room resonance and subtle machine sounds integrated into the track. No captions, titles, logos, labels, watermarks, extra dialogue, or visible writing.",
            "width": 1920,
            "height": 1080,
            "duration": 12
        })


asyncio.run(main())
curl https://api.runware.ai/v1 \
  -H "Authorization: Bearer $RUNWARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '[
    {
      "taskType": "videoInference",
      "taskUUID": "d263055d-29cc-46e3-8971-a2631caa3178",
      "model": "openai:3@2",
      "positivePrompt": "Create a finished 12-second, 16:9 music-video segment for an original art-pop single. Inside a vast, immaculate agricultural seed archive, long stainless-steel sorting tables and analog conveyor belts form strong geometric lines beneath cream-colored fluorescent ceiling panels. Thousands of dry beans and corn kernels in rust, mustard, ivory, black, and dusty blue travel across the machinery. The lead singer is a poised woman in her early thirties wearing a sharply tailored cobalt work jacket, ochre shirt, and dark trousers; four seed technicians in coordinated off-white uniforms work behind her.\n\nOne continuous choreographed shot with no cuts. Begin in a symmetrical wide shot looking straight down the central aisle as the singer walks toward camera. The camera smoothly dollies backward at her pace with subtle 35mm handheld texture. On each beat, the technicians perform crisp, believable sorting gestures—tilting steel trays, sliding archival drawers, and releasing streams of seeds that create graphic waves across the tables. At four seconds, the singer reaches foreground and sings directly into the lens with accurate, natural lip synchronization: “Count every kernel, keep a little summer, sealed beneath the mountain till the green returns.” During “little summer,” the camera makes a gentle half-orbit around her while a fan lifts the corner of her jacket; the moving seed colors create rhythmic bands in the background. End on a medium close-up as she holds eye contact and the conveyors stop together on the final word.\n\nVisual direction: 1970s agricultural educational film reimagined as refined contemporary fashion cinema; clean production design, restrained deadpan choreography, warm fine-grain 16mm texture, slightly faded cream highlights, rich cobalt and ochre accents, realistic skin and materials, soft halation around practical lights. Controlled, confident, quietly hopeful mood. Original synchronized audio: dry analog drum-machine groove at 112 BPM, rubbery mono synth bass, muted handclaps, small metallic conveyor percussion, and an intimate clear female alto singing the specified lyric. Natural room resonance and subtle machine sounds integrated into the track. No captions, titles, logos, labels, watermarks, extra dialogue, or visible writing.",
      "width": 1920,
      "height": 1080,
      "duration": 12
    }
  ]'
runware run openai:3@2 \
  positivePrompt="Create a finished 12-second, 16:9 music-video segment for an original art-pop single. Inside a vast, immaculate agricultural seed archive, long stainless-steel sorting tables and analog conveyor belts form strong geometric lines beneath cream-colored fluorescent ceiling panels. Thousands of dry beans and corn kernels in rust, mustard, ivory, black, and dusty blue travel across the machinery. The lead singer is a poised woman in her early thirties wearing a sharply tailored cobalt work jacket, ochre shirt, and dark trousers; four seed technicians in coordinated off-white uniforms work behind her.

One continuous choreographed shot with no cuts. Begin in a symmetrical wide shot looking straight down the central aisle as the singer walks toward camera. The camera smoothly dollies backward at her pace with subtle 35mm handheld texture. On each beat, the technicians perform crisp, believable sorting gestures—tilting steel trays, sliding archival drawers, and releasing streams of seeds that create graphic waves across the tables. At four seconds, the singer reaches foreground and sings directly into the lens with accurate, natural lip synchronization: “Count every kernel, keep a little summer, sealed beneath the mountain till the green returns.” During “little summer,” the camera makes a gentle half-orbit around her while a fan lifts the corner of her jacket; the moving seed colors create rhythmic bands in the background. End on a medium close-up as she holds eye contact and the conveyors stop together on the final word.

Visual direction: 1970s agricultural educational film reimagined as refined contemporary fashion cinema; clean production design, restrained deadpan choreography, warm fine-grain 16mm texture, slightly faded cream highlights, rich cobalt and ochre accents, realistic skin and materials, soft halation around practical lights. Controlled, confident, quietly hopeful mood. Original synchronized audio: dry analog drum-machine groove at 112 BPM, rubbery mono synth bass, muted handclaps, small metallic conveyor percussion, and an intimate clear female alto singing the specified lyric. Natural room resonance and subtle machine sounds integrated into the track. No captions, titles, logos, labels, watermarks, extra dialogue, or visible writing." \
  width=1920 \
  height=1080 \
  duration=12
{
  "taskType": "videoInference",
  "taskUUID": "d263055d-29cc-46e3-8971-a2631caa3178",
  "model": "openai:3@2",
  "positivePrompt": "Create a finished 12-second, 16:9 music-video segment for an original art-pop single. Inside a vast, immaculate agricultural seed archive, long stainless-steel sorting tables and analog conveyor belts form strong geometric lines beneath cream-colored fluorescent ceiling panels. Thousands of dry beans and corn kernels in rust, mustard, ivory, black, and dusty blue travel across the machinery. The lead singer is a poised woman in her early thirties wearing a sharply tailored cobalt work jacket, ochre shirt, and dark trousers; four seed technicians in coordinated off-white uniforms work behind her.\n\nOne continuous choreographed shot with no cuts. Begin in a symmetrical wide shot looking straight down the central aisle as the singer walks toward camera. The camera smoothly dollies backward at her pace with subtle 35mm handheld texture. On each beat, the technicians perform crisp, believable sorting gestures—tilting steel trays, sliding archival drawers, and releasing streams of seeds that create graphic waves across the tables. At four seconds, the singer reaches foreground and sings directly into the lens with accurate, natural lip synchronization: “Count every kernel, keep a little summer, sealed beneath the mountain till the green returns.” During “little summer,” the camera makes a gentle half-orbit around her while a fan lifts the corner of her jacket; the moving seed colors create rhythmic bands in the background. End on a medium close-up as she holds eye contact and the conveyors stop together on the final word.\n\nVisual direction: 1970s agricultural educational film reimagined as refined contemporary fashion cinema; clean production design, restrained deadpan choreography, warm fine-grain 16mm texture, slightly faded cream highlights, rich cobalt and ochre accents, realistic skin and materials, soft halation around practical lights. Controlled, confident, quietly hopeful mood. Original synchronized audio: dry analog drum-machine groove at 112 BPM, rubbery mono synth bass, muted handclaps, small metallic conveyor percussion, and an intimate clear female alto singing the specified lyric. Natural room resonance and subtle machine sounds integrated into the track. No captions, titles, logos, labels, watermarks, extra dialogue, or visible writing.",
  "width": 1920,
  "height": 1080,
  "duration": 12
}
Response
{
  "taskType": "videoInference",
  "taskUUID": "d263055d-29cc-46e3-8971-a2631caa3178",
  "videoUUID": "c4749985-9e4f-4c73-834a-44bffba734ff",
  "videoURL": "https://vm.runware.ai/video/os/a01d21/ws/5/vi/c4749985-9e4f-4c73-834a-44bffba734ff.mp4",
  "seed": 946930883,
  "cost": 8.4,
  "outputs": {
    "videoId": "video_6a960a4f761081939fa47a82de362fcd0b4334a6b3e0037c"
  }
}
first-frame

Tailored Beekeeper Suit Fashion Reel

import { createClient } from '@runware/sdk'

const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()

const [result] = await client.run({
  model: 'openai:3@2',
  positivePrompt: 'Continue directly from the supplied first frame as an 8-second vertical luxury workwear fashion reel. Preserve the same adult model, tailored ivory beekeeping suit, mustard leather gloves, black mesh veil, copper smoker, hive arrangement, warm sunlight, and restrained cream-and-marigold palette throughout. Do not redesign or replace any garment details.\n\n0–2 seconds: begin with the exact opening composition. The model takes one deliberate step toward camera, creating natural folds through the waxed cotton. She raises the sculptural mesh veil smoothly over her face while a light breeze moves the fabric. The camera makes a subtle stabilized push inward.\n\n2–5 seconds: transition into a smooth three-quarter orbit as the model pivots beside the hives. Keep her anatomy and outfit temporally consistent. Show the articulated elbow, reinforced knee, structured hip pocket, metal zipper, and long protective gauntlet in crisp editorial detail. A few bees drift through the sunbeam without swarming or obscuring the clothing.\n\n5–8 seconds: settle into a confident waist-up hero framing. The model closes one glove cuff, then gently compresses the smoker bellows. One elegant ribbon of pale smoke curls across the backlight while she looks calmly past camera. End on a composed, premium fashion pose with the veil and tailoring clearly readable.\n\nNaturalistic motion, refined contemporary fashion cinematography, realistic fabric physics, restrained choreography, smooth gimbal movement, shallow but stable depth of field, warm hard sunlight with clean shadows, subtle 35mm film texture. Synchronized sound: soft boot step, waxed-cotton rustle, zipper hardware, quiet bee hum, one gentle smoker-bellows breath, and a minimal dry percussive pulse. No speech, captions, lettering, logos, cuts to unrelated scenes, extra people, or product substitutions.',
  width: 1080,
  height: 1920,
  duration: 8,
  inputs: {
    frameImages: [
      'https://assets.runware.ai/assets/inputs/bc3ca0fa-437b-4cf3-ba8b-4edacc3c7471.jpg'
    ]
  }
})
import asyncio
import os

from runware import Runware


async def main():
    async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
        results = await client.run({
            "model": "openai:3@2",
            "positivePrompt": "Continue directly from the supplied first frame as an 8-second vertical luxury workwear fashion reel. Preserve the same adult model, tailored ivory beekeeping suit, mustard leather gloves, black mesh veil, copper smoker, hive arrangement, warm sunlight, and restrained cream-and-marigold palette throughout. Do not redesign or replace any garment details.\n\n0–2 seconds: begin with the exact opening composition. The model takes one deliberate step toward camera, creating natural folds through the waxed cotton. She raises the sculptural mesh veil smoothly over her face while a light breeze moves the fabric. The camera makes a subtle stabilized push inward.\n\n2–5 seconds: transition into a smooth three-quarter orbit as the model pivots beside the hives. Keep her anatomy and outfit temporally consistent. Show the articulated elbow, reinforced knee, structured hip pocket, metal zipper, and long protective gauntlet in crisp editorial detail. A few bees drift through the sunbeam without swarming or obscuring the clothing.\n\n5–8 seconds: settle into a confident waist-up hero framing. The model closes one glove cuff, then gently compresses the smoker bellows. One elegant ribbon of pale smoke curls across the backlight while she looks calmly past camera. End on a composed, premium fashion pose with the veil and tailoring clearly readable.\n\nNaturalistic motion, refined contemporary fashion cinematography, realistic fabric physics, restrained choreography, smooth gimbal movement, shallow but stable depth of field, warm hard sunlight with clean shadows, subtle 35mm film texture. Synchronized sound: soft boot step, waxed-cotton rustle, zipper hardware, quiet bee hum, one gentle smoker-bellows breath, and a minimal dry percussive pulse. No speech, captions, lettering, logos, cuts to unrelated scenes, extra people, or product substitutions.",
            "width": 1080,
            "height": 1920,
            "duration": 8,
            "inputs": {
                "frameImages": [
                    "https://assets.runware.ai/assets/inputs/bc3ca0fa-437b-4cf3-ba8b-4edacc3c7471.jpg"
                ]
            }
        })


asyncio.run(main())
curl https://api.runware.ai/v1 \
  -H "Authorization: Bearer $RUNWARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '[
    {
      "taskType": "videoInference",
      "taskUUID": "bed421b1-42b2-4aa8-ac48-64d157ff6862",
      "model": "openai:3@2",
      "positivePrompt": "Continue directly from the supplied first frame as an 8-second vertical luxury workwear fashion reel. Preserve the same adult model, tailored ivory beekeeping suit, mustard leather gloves, black mesh veil, copper smoker, hive arrangement, warm sunlight, and restrained cream-and-marigold palette throughout. Do not redesign or replace any garment details.\n\n0–2 seconds: begin with the exact opening composition. The model takes one deliberate step toward camera, creating natural folds through the waxed cotton. She raises the sculptural mesh veil smoothly over her face while a light breeze moves the fabric. The camera makes a subtle stabilized push inward.\n\n2–5 seconds: transition into a smooth three-quarter orbit as the model pivots beside the hives. Keep her anatomy and outfit temporally consistent. Show the articulated elbow, reinforced knee, structured hip pocket, metal zipper, and long protective gauntlet in crisp editorial detail. A few bees drift through the sunbeam without swarming or obscuring the clothing.\n\n5–8 seconds: settle into a confident waist-up hero framing. The model closes one glove cuff, then gently compresses the smoker bellows. One elegant ribbon of pale smoke curls across the backlight while she looks calmly past camera. End on a composed, premium fashion pose with the veil and tailoring clearly readable.\n\nNaturalistic motion, refined contemporary fashion cinematography, realistic fabric physics, restrained choreography, smooth gimbal movement, shallow but stable depth of field, warm hard sunlight with clean shadows, subtle 35mm film texture. Synchronized sound: soft boot step, waxed-cotton rustle, zipper hardware, quiet bee hum, one gentle smoker-bellows breath, and a minimal dry percussive pulse. No speech, captions, lettering, logos, cuts to unrelated scenes, extra people, or product substitutions.",
      "width": 1080,
      "height": 1920,
      "duration": 8,
      "inputs": {
        "frameImages": [
          "https://assets.runware.ai/assets/inputs/bc3ca0fa-437b-4cf3-ba8b-4edacc3c7471.jpg"
        ]
      }
    }
  ]'
runware run openai:3@2 \
  positivePrompt="Continue directly from the supplied first frame as an 8-second vertical luxury workwear fashion reel. Preserve the same adult model, tailored ivory beekeeping suit, mustard leather gloves, black mesh veil, copper smoker, hive arrangement, warm sunlight, and restrained cream-and-marigold palette throughout. Do not redesign or replace any garment details.

0–2 seconds: begin with the exact opening composition. The model takes one deliberate step toward camera, creating natural folds through the waxed cotton. She raises the sculptural mesh veil smoothly over her face while a light breeze moves the fabric. The camera makes a subtle stabilized push inward.

2–5 seconds: transition into a smooth three-quarter orbit as the model pivots beside the hives. Keep her anatomy and outfit temporally consistent. Show the articulated elbow, reinforced knee, structured hip pocket, metal zipper, and long protective gauntlet in crisp editorial detail. A few bees drift through the sunbeam without swarming or obscuring the clothing.

5–8 seconds: settle into a confident waist-up hero framing. The model closes one glove cuff, then gently compresses the smoker bellows. One elegant ribbon of pale smoke curls across the backlight while she looks calmly past camera. End on a composed, premium fashion pose with the veil and tailoring clearly readable.

Naturalistic motion, refined contemporary fashion cinematography, realistic fabric physics, restrained choreography, smooth gimbal movement, shallow but stable depth of field, warm hard sunlight with clean shadows, subtle 35mm film texture. Synchronized sound: soft boot step, waxed-cotton rustle, zipper hardware, quiet bee hum, one gentle smoker-bellows breath, and a minimal dry percussive pulse. No speech, captions, lettering, logos, cuts to unrelated scenes, extra people, or product substitutions." \
  width=1080 \
  height=1920 \
  duration=8 \
  inputs.frameImages.0=https://assets.runware.ai/assets/inputs/bc3ca0fa-437b-4cf3-ba8b-4edacc3c7471.jpg
{
  "taskType": "videoInference",
  "taskUUID": "bed421b1-42b2-4aa8-ac48-64d157ff6862",
  "model": "openai:3@2",
  "positivePrompt": "Continue directly from the supplied first frame as an 8-second vertical luxury workwear fashion reel. Preserve the same adult model, tailored ivory beekeeping suit, mustard leather gloves, black mesh veil, copper smoker, hive arrangement, warm sunlight, and restrained cream-and-marigold palette throughout. Do not redesign or replace any garment details.\n\n0–2 seconds: begin with the exact opening composition. The model takes one deliberate step toward camera, creating natural folds through the waxed cotton. She raises the sculptural mesh veil smoothly over her face while a light breeze moves the fabric. The camera makes a subtle stabilized push inward.\n\n2–5 seconds: transition into a smooth three-quarter orbit as the model pivots beside the hives. Keep her anatomy and outfit temporally consistent. Show the articulated elbow, reinforced knee, structured hip pocket, metal zipper, and long protective gauntlet in crisp editorial detail. A few bees drift through the sunbeam without swarming or obscuring the clothing.\n\n5–8 seconds: settle into a confident waist-up hero framing. The model closes one glove cuff, then gently compresses the smoker bellows. One elegant ribbon of pale smoke curls across the backlight while she looks calmly past camera. End on a composed, premium fashion pose with the veil and tailoring clearly readable.\n\nNaturalistic motion, refined contemporary fashion cinematography, realistic fabric physics, restrained choreography, smooth gimbal movement, shallow but stable depth of field, warm hard sunlight with clean shadows, subtle 35mm film texture. Synchronized sound: soft boot step, waxed-cotton rustle, zipper hardware, quiet bee hum, one gentle smoker-bellows breath, and a minimal dry percussive pulse. No speech, captions, lettering, logos, cuts to unrelated scenes, extra people, or product substitutions.",
  "width": 1080,
  "height": 1920,
  "duration": 8,
  "inputs": {
    "frameImages": [
      "https://assets.runware.ai/assets/inputs/bc3ca0fa-437b-4cf3-ba8b-4edacc3c7471.jpg"
    ]
  }
}
Response
{
  "taskType": "videoInference",
  "taskUUID": "bed421b1-42b2-4aa8-ac48-64d157ff6862",
  "videoUUID": "6cb99c8f-3e9b-486f-b800-8cf514066b90",
  "videoURL": "https://vm.runware.ai/video/os/a09dlim3/ws/5/vi/6cb99c8f-3e9b-486f-b800-8cf514066b90.mp4",
  "seed": 1584436543,
  "cost": 5.6,
  "outputs": {
    "videoId": "video_6a96440a23f48191ae7d0b6f9767570c029036f3d50037a4"
  }
}
Text to Video

Bioluminescent Mangrove Monsoon Twilight

import { createClient } from '@runware/sdk'

const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()

const [result] = await client.run({
  model: 'openai:3@2',
  positivePrompt: 'A cinematic wide shot drifting slowly through a flooded mangrove forest at twilight during a warm monsoon rain. The tangled roots rise from mirror-dark water, and every raindrop disturbs glowing turquoise bioluminescent algae that ripple outward in delicate rings. Small lantern-like jelly insects hover between branches, casting soft teal and amber light. A narrow wooden skiff sits half-hidden among the roots, covered in wet woven baskets, seashell charms, and a dim oil lamp. In the distance, thunderclouds pulse with silent violet lightning behind layers of mist. The camera glides forward at water level with subtle parallax through leaves and hanging vines, revealing tiny tree frogs, drifting pollen, and rain sliding from glossy foliage. Soundscape: dense rainfall on leaves and water, soft insect chorus, distant thunder, creaking wood, gentle water lapping. Ultra-detailed, naturalistic cinematic fantasy, precise reflections, volumetric mist, moody color contrast, believable physics, cohesive atmosphere, premium production quality.',
  width: 1792,
  height: 1024,
  duration: 8
})
import asyncio
import os

from runware import Runware


async def main():
    async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
        results = await client.run({
            "model": "openai:3@2",
            "positivePrompt": "A cinematic wide shot drifting slowly through a flooded mangrove forest at twilight during a warm monsoon rain. The tangled roots rise from mirror-dark water, and every raindrop disturbs glowing turquoise bioluminescent algae that ripple outward in delicate rings. Small lantern-like jelly insects hover between branches, casting soft teal and amber light. A narrow wooden skiff sits half-hidden among the roots, covered in wet woven baskets, seashell charms, and a dim oil lamp. In the distance, thunderclouds pulse with silent violet lightning behind layers of mist. The camera glides forward at water level with subtle parallax through leaves and hanging vines, revealing tiny tree frogs, drifting pollen, and rain sliding from glossy foliage. Soundscape: dense rainfall on leaves and water, soft insect chorus, distant thunder, creaking wood, gentle water lapping. Ultra-detailed, naturalistic cinematic fantasy, precise reflections, volumetric mist, moody color contrast, believable physics, cohesive atmosphere, premium production quality.",
            "width": 1792,
            "height": 1024,
            "duration": 8
        })


asyncio.run(main())
curl https://api.runware.ai/v1 \
  -H "Authorization: Bearer $RUNWARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '[
    {
      "taskType": "videoInference",
      "taskUUID": "982c01aa-813a-4434-aa42-3e512808fd94",
      "model": "openai:3@2",
      "positivePrompt": "A cinematic wide shot drifting slowly through a flooded mangrove forest at twilight during a warm monsoon rain. The tangled roots rise from mirror-dark water, and every raindrop disturbs glowing turquoise bioluminescent algae that ripple outward in delicate rings. Small lantern-like jelly insects hover between branches, casting soft teal and amber light. A narrow wooden skiff sits half-hidden among the roots, covered in wet woven baskets, seashell charms, and a dim oil lamp. In the distance, thunderclouds pulse with silent violet lightning behind layers of mist. The camera glides forward at water level with subtle parallax through leaves and hanging vines, revealing tiny tree frogs, drifting pollen, and rain sliding from glossy foliage. Soundscape: dense rainfall on leaves and water, soft insect chorus, distant thunder, creaking wood, gentle water lapping. Ultra-detailed, naturalistic cinematic fantasy, precise reflections, volumetric mist, moody color contrast, believable physics, cohesive atmosphere, premium production quality.",
      "width": 1792,
      "height": 1024,
      "duration": 8
    }
  ]'
runware run openai:3@2 \
  positivePrompt="A cinematic wide shot drifting slowly through a flooded mangrove forest at twilight during a warm monsoon rain. The tangled roots rise from mirror-dark water, and every raindrop disturbs glowing turquoise bioluminescent algae that ripple outward in delicate rings. Small lantern-like jelly insects hover between branches, casting soft teal and amber light. A narrow wooden skiff sits half-hidden among the roots, covered in wet woven baskets, seashell charms, and a dim oil lamp. In the distance, thunderclouds pulse with silent violet lightning behind layers of mist. The camera glides forward at water level with subtle parallax through leaves and hanging vines, revealing tiny tree frogs, drifting pollen, and rain sliding from glossy foliage. Soundscape: dense rainfall on leaves and water, soft insect chorus, distant thunder, creaking wood, gentle water lapping. Ultra-detailed, naturalistic cinematic fantasy, precise reflections, volumetric mist, moody color contrast, believable physics, cohesive atmosphere, premium production quality." \
  width=1792 \
  height=1024 \
  duration=8
{
  "taskType": "videoInference",
  "taskUUID": "982c01aa-813a-4434-aa42-3e512808fd94",
  "model": "openai:3@2",
  "positivePrompt": "A cinematic wide shot drifting slowly through a flooded mangrove forest at twilight during a warm monsoon rain. The tangled roots rise from mirror-dark water, and every raindrop disturbs glowing turquoise bioluminescent algae that ripple outward in delicate rings. Small lantern-like jelly insects hover between branches, casting soft teal and amber light. A narrow wooden skiff sits half-hidden among the roots, covered in wet woven baskets, seashell charms, and a dim oil lamp. In the distance, thunderclouds pulse with silent violet lightning behind layers of mist. The camera glides forward at water level with subtle parallax through leaves and hanging vines, revealing tiny tree frogs, drifting pollen, and rain sliding from glossy foliage. Soundscape: dense rainfall on leaves and water, soft insect chorus, distant thunder, creaking wood, gentle water lapping. Ultra-detailed, naturalistic cinematic fantasy, precise reflections, volumetric mist, moody color contrast, believable physics, cohesive atmosphere, premium production quality.",
  "width": 1792,
  "height": 1024,
  "duration": 8
}
Response
{
  "taskType": "videoInference",
  "taskUUID": "982c01aa-813a-4434-aa42-3e512808fd94",
  "videoUUID": "93666777-2328-4441-a0d4-b6df40eb0a1a",
  "videoURL": "https://vm.runware.ai/video/os/a04d20/ws/5/vi/93666777-2328-4441-a0d4-b6df40eb0a1a.mp4",
  "seed": 443404476,
  "cost": 4,
  "outputs": {
    "videoId": "video_69c581a1a7688190bf24a177278b6187000b0b6342cde7b3"
  }
}
Text to Video

Bioluminescent Monsoon Market Night

import { createClient } from '@runware/sdk'

const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()

const [result] = await client.run({
  model: 'openai:3@2',
  positivePrompt: 'A cinematic nighttime scene in a flooded bioluminescent market built on stilts above a jungle river during warm monsoon rain. Glowing lantern-fruit hang from woven awnings, reflecting cyan, amber, and magenta light across rippling water. Vendors in translucent rain capes sell luminous shells, steaming tea, and strange orchids that pulse softly like living neon. The camera begins with a low glide just above the rain-dappled water, passing floating baskets and drifting petals, then rises into the market lane as pedestrians step across wooden planks and suspension bridges. A child pauses to launch a tiny paper boat with a candle into the current. In the distance, a colossal manta-shaped air barge drifts through mist between giant trees, its engine hum mixing with thunder. Sound design: layered rainfall on fabric and wood, water sloshing underfoot, soft market chatter in the background, tea kettles hissing, insects singing from the jungle, occasional thunder rolls, subtle rope creaks, gentle splashes, and the low passing drone of the air barge overhead. Visual style: ultra-detailed, premium cinematic realism with fantastical worldbuilding, natural human motion, wet surfaces, volumetric mist, expressive reflections, rich color separation, shallow depth of field in close passes, smooth camera movement, physically plausible lighting, and consistent character continuity throughout.',
  width: 1792,
  height: 1024,
  duration: 8
})
import asyncio
import os

from runware import Runware


async def main():
    async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
        results = await client.run({
            "model": "openai:3@2",
            "positivePrompt": "A cinematic nighttime scene in a flooded bioluminescent market built on stilts above a jungle river during warm monsoon rain. Glowing lantern-fruit hang from woven awnings, reflecting cyan, amber, and magenta light across rippling water. Vendors in translucent rain capes sell luminous shells, steaming tea, and strange orchids that pulse softly like living neon. The camera begins with a low glide just above the rain-dappled water, passing floating baskets and drifting petals, then rises into the market lane as pedestrians step across wooden planks and suspension bridges. A child pauses to launch a tiny paper boat with a candle into the current. In the distance, a colossal manta-shaped air barge drifts through mist between giant trees, its engine hum mixing with thunder. Sound design: layered rainfall on fabric and wood, water sloshing underfoot, soft market chatter in the background, tea kettles hissing, insects singing from the jungle, occasional thunder rolls, subtle rope creaks, gentle splashes, and the low passing drone of the air barge overhead. Visual style: ultra-detailed, premium cinematic realism with fantastical worldbuilding, natural human motion, wet surfaces, volumetric mist, expressive reflections, rich color separation, shallow depth of field in close passes, smooth camera movement, physically plausible lighting, and consistent character continuity throughout.",
            "width": 1792,
            "height": 1024,
            "duration": 8
        })


asyncio.run(main())
curl https://api.runware.ai/v1 \
  -H "Authorization: Bearer $RUNWARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '[
    {
      "taskType": "videoInference",
      "taskUUID": "a317f802-2c2a-4ef9-9187-745116d6ea2f",
      "model": "openai:3@2",
      "positivePrompt": "A cinematic nighttime scene in a flooded bioluminescent market built on stilts above a jungle river during warm monsoon rain. Glowing lantern-fruit hang from woven awnings, reflecting cyan, amber, and magenta light across rippling water. Vendors in translucent rain capes sell luminous shells, steaming tea, and strange orchids that pulse softly like living neon. The camera begins with a low glide just above the rain-dappled water, passing floating baskets and drifting petals, then rises into the market lane as pedestrians step across wooden planks and suspension bridges. A child pauses to launch a tiny paper boat with a candle into the current. In the distance, a colossal manta-shaped air barge drifts through mist between giant trees, its engine hum mixing with thunder. Sound design: layered rainfall on fabric and wood, water sloshing underfoot, soft market chatter in the background, tea kettles hissing, insects singing from the jungle, occasional thunder rolls, subtle rope creaks, gentle splashes, and the low passing drone of the air barge overhead. Visual style: ultra-detailed, premium cinematic realism with fantastical worldbuilding, natural human motion, wet surfaces, volumetric mist, expressive reflections, rich color separation, shallow depth of field in close passes, smooth camera movement, physically plausible lighting, and consistent character continuity throughout.",
      "width": 1792,
      "height": 1024,
      "duration": 8
    }
  ]'
runware run openai:3@2 \
  positivePrompt="A cinematic nighttime scene in a flooded bioluminescent market built on stilts above a jungle river during warm monsoon rain. Glowing lantern-fruit hang from woven awnings, reflecting cyan, amber, and magenta light across rippling water. Vendors in translucent rain capes sell luminous shells, steaming tea, and strange orchids that pulse softly like living neon. The camera begins with a low glide just above the rain-dappled water, passing floating baskets and drifting petals, then rises into the market lane as pedestrians step across wooden planks and suspension bridges. A child pauses to launch a tiny paper boat with a candle into the current. In the distance, a colossal manta-shaped air barge drifts through mist between giant trees, its engine hum mixing with thunder. Sound design: layered rainfall on fabric and wood, water sloshing underfoot, soft market chatter in the background, tea kettles hissing, insects singing from the jungle, occasional thunder rolls, subtle rope creaks, gentle splashes, and the low passing drone of the air barge overhead. Visual style: ultra-detailed, premium cinematic realism with fantastical worldbuilding, natural human motion, wet surfaces, volumetric mist, expressive reflections, rich color separation, shallow depth of field in close passes, smooth camera movement, physically plausible lighting, and consistent character continuity throughout." \
  width=1792 \
  height=1024 \
  duration=8
{
  "taskType": "videoInference",
  "taskUUID": "a317f802-2c2a-4ef9-9187-745116d6ea2f",
  "model": "openai:3@2",
  "positivePrompt": "A cinematic nighttime scene in a flooded bioluminescent market built on stilts above a jungle river during warm monsoon rain. Glowing lantern-fruit hang from woven awnings, reflecting cyan, amber, and magenta light across rippling water. Vendors in translucent rain capes sell luminous shells, steaming tea, and strange orchids that pulse softly like living neon. The camera begins with a low glide just above the rain-dappled water, passing floating baskets and drifting petals, then rises into the market lane as pedestrians step across wooden planks and suspension bridges. A child pauses to launch a tiny paper boat with a candle into the current. In the distance, a colossal manta-shaped air barge drifts through mist between giant trees, its engine hum mixing with thunder. Sound design: layered rainfall on fabric and wood, water sloshing underfoot, soft market chatter in the background, tea kettles hissing, insects singing from the jungle, occasional thunder rolls, subtle rope creaks, gentle splashes, and the low passing drone of the air barge overhead. Visual style: ultra-detailed, premium cinematic realism with fantastical worldbuilding, natural human motion, wet surfaces, volumetric mist, expressive reflections, rich color separation, shallow depth of field in close passes, smooth camera movement, physically plausible lighting, and consistent character continuity throughout.",
  "width": 1792,
  "height": 1024,
  "duration": 8
}
Response
{
  "taskType": "videoInference",
  "taskUUID": "a317f802-2c2a-4ef9-9187-745116d6ea2f",
  "videoUUID": "abead010-c78e-4d93-b69d-1cbffd05b2cf",
  "videoURL": "https://vm.runware.ai/video/os/a04d20/ws/5/vi/abead010-c78e-4d93-b69d-1cbffd05b2cf.mp4",
  "seed": 1222567127,
  "cost": 4,
  "outputs": {
    "videoId": "video_69c581a64cc88190b947c1105355cc0d072f1184959760cd"
  }
}
Text to Video

Bioluminescent Monsoon Market Alley

import { createClient } from '@runware/sdk'

const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()

const [result] = await client.run({
  model: 'openai:3@2',
  positivePrompt: 'A cinematic night market alley during a warm monsoon rain in a fantastical coastal megacity, glowing bioluminescent lanterns hanging between buildings, puddles reflecting teal, amber, and magenta light, steam rising from street food stalls, silk awnings fluttering in the wind, shoppers with translucent umbrellas, a quiet stray cat weaving between crates, intricate signage in an invented script, ultra-detailed wet cobblestones, subtle handheld camera dolly forward through the alley, natural parallax, realistic rain interaction on fabric and metal, occasional distant tram passing at the end of the street, immersive synchronized soundscape of rainfall, soft market chatter, sizzling food, glass clinks, rolling thunder, and humming neon. Mood: dreamy, intimate, alive, premium cinematic realism with poetic color grading and excellent continuity.',
  width: 1792,
  height: 1024,
  duration: 8
})
import asyncio
import os

from runware import Runware


async def main():
    async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
        results = await client.run({
            "model": "openai:3@2",
            "positivePrompt": "A cinematic night market alley during a warm monsoon rain in a fantastical coastal megacity, glowing bioluminescent lanterns hanging between buildings, puddles reflecting teal, amber, and magenta light, steam rising from street food stalls, silk awnings fluttering in the wind, shoppers with translucent umbrellas, a quiet stray cat weaving between crates, intricate signage in an invented script, ultra-detailed wet cobblestones, subtle handheld camera dolly forward through the alley, natural parallax, realistic rain interaction on fabric and metal, occasional distant tram passing at the end of the street, immersive synchronized soundscape of rainfall, soft market chatter, sizzling food, glass clinks, rolling thunder, and humming neon. Mood: dreamy, intimate, alive, premium cinematic realism with poetic color grading and excellent continuity.",
            "width": 1792,
            "height": 1024,
            "duration": 8
        })


asyncio.run(main())
curl https://api.runware.ai/v1 \
  -H "Authorization: Bearer $RUNWARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '[
    {
      "taskType": "videoInference",
      "taskUUID": "7da8ad5f-5b6c-475e-b2f7-bd5a47e90170",
      "model": "openai:3@2",
      "positivePrompt": "A cinematic night market alley during a warm monsoon rain in a fantastical coastal megacity, glowing bioluminescent lanterns hanging between buildings, puddles reflecting teal, amber, and magenta light, steam rising from street food stalls, silk awnings fluttering in the wind, shoppers with translucent umbrellas, a quiet stray cat weaving between crates, intricate signage in an invented script, ultra-detailed wet cobblestones, subtle handheld camera dolly forward through the alley, natural parallax, realistic rain interaction on fabric and metal, occasional distant tram passing at the end of the street, immersive synchronized soundscape of rainfall, soft market chatter, sizzling food, glass clinks, rolling thunder, and humming neon. Mood: dreamy, intimate, alive, premium cinematic realism with poetic color grading and excellent continuity.",
      "width": 1792,
      "height": 1024,
      "duration": 8
    }
  ]'
runware run openai:3@2 \
  positivePrompt="A cinematic night market alley during a warm monsoon rain in a fantastical coastal megacity, glowing bioluminescent lanterns hanging between buildings, puddles reflecting teal, amber, and magenta light, steam rising from street food stalls, silk awnings fluttering in the wind, shoppers with translucent umbrellas, a quiet stray cat weaving between crates, intricate signage in an invented script, ultra-detailed wet cobblestones, subtle handheld camera dolly forward through the alley, natural parallax, realistic rain interaction on fabric and metal, occasional distant tram passing at the end of the street, immersive synchronized soundscape of rainfall, soft market chatter, sizzling food, glass clinks, rolling thunder, and humming neon. Mood: dreamy, intimate, alive, premium cinematic realism with poetic color grading and excellent continuity." \
  width=1792 \
  height=1024 \
  duration=8
{
  "taskType": "videoInference",
  "taskUUID": "7da8ad5f-5b6c-475e-b2f7-bd5a47e90170",
  "model": "openai:3@2",
  "positivePrompt": "A cinematic night market alley during a warm monsoon rain in a fantastical coastal megacity, glowing bioluminescent lanterns hanging between buildings, puddles reflecting teal, amber, and magenta light, steam rising from street food stalls, silk awnings fluttering in the wind, shoppers with translucent umbrellas, a quiet stray cat weaving between crates, intricate signage in an invented script, ultra-detailed wet cobblestones, subtle handheld camera dolly forward through the alley, natural parallax, realistic rain interaction on fabric and metal, occasional distant tram passing at the end of the street, immersive synchronized soundscape of rainfall, soft market chatter, sizzling food, glass clinks, rolling thunder, and humming neon. Mood: dreamy, intimate, alive, premium cinematic realism with poetic color grading and excellent continuity.",
  "width": 1792,
  "height": 1024,
  "duration": 8
}
Response
{
  "taskType": "videoInference",
  "taskUUID": "7da8ad5f-5b6c-475e-b2f7-bd5a47e90170",
  "videoUUID": "f850f5f0-c1dc-499c-954d-2191221a2dfe",
  "videoURL": "https://vm.runware.ai/video/os/a13d12/ws/5/vi/f850f5f0-c1dc-499c-954d-2191221a2dfe.mp4",
  "seed": 1846504039,
  "cost": 4,
  "outputs": {
    "videoId": "video_69c5818f06688190ba8588b03adaca6f0cdcb508961f2595"
  }
}
Image to Video

Moonlit Festival Rooftop Tableau

import { createClient } from '@runware/sdk'

const client = await createClient({ apiKey: process.env.RUNWARE_API_KEY })
await client.connect()

const [result] = await client.run({
  model: 'openai:3@2',
  positivePrompt: 'Using the provided first-frame image as the opening shot, create a cinematic night sequence with synchronized ambient festival sound: soft wind, distant drums, muffled crowd chatter, lantern paper rustling, river water below. The camera begins close on the rooftop tableau, then slowly glides sideways and slightly forward as lanterns drift into the sky. The woman gently rises, steadies the fox mask in one hand, and turns toward the glowing city. Her coat and hair move naturally in the breeze. A few sparks from fireworks reflect in the brass radio and wet roof tiles. Keep the original character identity, costume details, rooftop props, lighting palette, and old-city architecture consistent with the first frame. Emphasize realistic motion, elegant parallax, volumetric moonlight, drifting smoke, rich texture, and strong temporal continuity. Premium cinematic look, subtle emotional wonder, highly polished production quality.',
  width: 1792,
  height: 1024,
  duration: 8,
  inputs: {
    frameImages: [
      {
        image: 'https://assets.runware.ai/assets/inputs/b66a1179-9061-4898-9a23-7a78b91eb578.jpg',
        frame: 'first'
      }
    ]
  }
})
import asyncio
import os

from runware import Runware


async def main():
    async with Runware(api_key=os.environ["RUNWARE_API_KEY"]) as client:
        results = await client.run({
            "model": "openai:3@2",
            "positivePrompt": "Using the provided first-frame image as the opening shot, create a cinematic night sequence with synchronized ambient festival sound: soft wind, distant drums, muffled crowd chatter, lantern paper rustling, river water below. The camera begins close on the rooftop tableau, then slowly glides sideways and slightly forward as lanterns drift into the sky. The woman gently rises, steadies the fox mask in one hand, and turns toward the glowing city. Her coat and hair move naturally in the breeze. A few sparks from fireworks reflect in the brass radio and wet roof tiles. Keep the original character identity, costume details, rooftop props, lighting palette, and old-city architecture consistent with the first frame. Emphasize realistic motion, elegant parallax, volumetric moonlight, drifting smoke, rich texture, and strong temporal continuity. Premium cinematic look, subtle emotional wonder, highly polished production quality.",
            "width": 1792,
            "height": 1024,
            "duration": 8,
            "inputs": {
                "frameImages": [
                    {
                        "image": "https://assets.runware.ai/assets/inputs/b66a1179-9061-4898-9a23-7a78b91eb578.jpg",
                        "frame": "first"
                    }
                ]
            }
        })


asyncio.run(main())
curl https://api.runware.ai/v1 \
  -H "Authorization: Bearer $RUNWARE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '[
    {
      "taskType": "videoInference",
      "taskUUID": "94754aa0-2571-468d-9795-7b4cf3c73f92",
      "model": "openai:3@2",
      "positivePrompt": "Using the provided first-frame image as the opening shot, create a cinematic night sequence with synchronized ambient festival sound: soft wind, distant drums, muffled crowd chatter, lantern paper rustling, river water below. The camera begins close on the rooftop tableau, then slowly glides sideways and slightly forward as lanterns drift into the sky. The woman gently rises, steadies the fox mask in one hand, and turns toward the glowing city. Her coat and hair move naturally in the breeze. A few sparks from fireworks reflect in the brass radio and wet roof tiles. Keep the original character identity, costume details, rooftop props, lighting palette, and old-city architecture consistent with the first frame. Emphasize realistic motion, elegant parallax, volumetric moonlight, drifting smoke, rich texture, and strong temporal continuity. Premium cinematic look, subtle emotional wonder, highly polished production quality.",
      "width": 1792,
      "height": 1024,
      "duration": 8,
      "inputs": {
        "frameImages": [
          {
            "image": "https://assets.runware.ai/assets/inputs/b66a1179-9061-4898-9a23-7a78b91eb578.jpg",
            "frame": "first"
          }
        ]
      }
    }
  ]'
runware run openai:3@2 \
  positivePrompt="Using the provided first-frame image as the opening shot, create a cinematic night sequence with synchronized ambient festival sound: soft wind, distant drums, muffled crowd chatter, lantern paper rustling, river water below. The camera begins close on the rooftop tableau, then slowly glides sideways and slightly forward as lanterns drift into the sky. The woman gently rises, steadies the fox mask in one hand, and turns toward the glowing city. Her coat and hair move naturally in the breeze. A few sparks from fireworks reflect in the brass radio and wet roof tiles. Keep the original character identity, costume details, rooftop props, lighting palette, and old-city architecture consistent with the first frame. Emphasize realistic motion, elegant parallax, volumetric moonlight, drifting smoke, rich texture, and strong temporal continuity. Premium cinematic look, subtle emotional wonder, highly polished production quality." \
  width=1792 \
  height=1024 \
  duration=8 \
  inputs.frameImages.0.image=https://assets.runware.ai/assets/inputs/b66a1179-9061-4898-9a23-7a78b91eb578.jpg \
  inputs.frameImages.0.frame=first
{
  "taskType": "videoInference",
  "taskUUID": "94754aa0-2571-468d-9795-7b4cf3c73f92",
  "model": "openai:3@2",
  "positivePrompt": "Using the provided first-frame image as the opening shot, create a cinematic night sequence with synchronized ambient festival sound: soft wind, distant drums, muffled crowd chatter, lantern paper rustling, river water below. The camera begins close on the rooftop tableau, then slowly glides sideways and slightly forward as lanterns drift into the sky. The woman gently rises, steadies the fox mask in one hand, and turns toward the glowing city. Her coat and hair move naturally in the breeze. A few sparks from fireworks reflect in the brass radio and wet roof tiles. Keep the original character identity, costume details, rooftop props, lighting palette, and old-city architecture consistent with the first frame. Emphasize realistic motion, elegant parallax, volumetric moonlight, drifting smoke, rich texture, and strong temporal continuity. Premium cinematic look, subtle emotional wonder, highly polished production quality.",
  "width": 1792,
  "height": 1024,
  "duration": 8,
  "inputs": {
    "frameImages": [
      {
        "image": "https://assets.runware.ai/assets/inputs/b66a1179-9061-4898-9a23-7a78b91eb578.jpg",
        "frame": "first"
      }
    ]
  }
}
Response
{
  "taskType": "videoInference",
  "taskUUID": "94754aa0-2571-468d-9795-7b4cf3c73f92",
  "videoUUID": "cfa4a4a0-fd39-43e6-949e-53ed9d074467",
  "videoURL": "https://vm.runware.ai/video/os/a04d20/ws/5/vi/cfa4a4a0-fd39-43e6-949e-53ed9d074467.mp4",
  "seed": 1166807891,
  "cost": 4,
  "outputs": {
    "videoId": "video_69c581a181ec8198a968097b09d5b9510b4afedd91e5377a"
  }
}