## Basic model info - Model name: bytedance/bytedance seedance-2.0 - Model description: Generate videos with audio using Seedance 2.0 with multimodal reference support. - Endpoint name: text-to-video ## Model schema The model schema is defined in the OpenAPI schema: [OpenAPI Schema](https://oapi.sunra.ai/main/bytedance/seedance-2.0/latest.json) ### Model input schema The model input schema is: ```json { "description": "Input model for Seedance 2.0 text-to-video generation.", "properties": { "prompt": { "description": "Text prompt for video generation. Supports Chinese (recommended <= 500 chars) and English (recommended <= 1000 words).", "title": "Prompt", "type": "string", "x-sr-order": 200 }, "duration": { "default": 5, "description": "Duration of the video in seconds (4-15, or -1 for auto). Must be >= 4 unless -1.", "maximum": 15, "minimum": -1, "title": "Duration", "type": "integer", "x-sr-order": 403 }, "generate_audio": { "default": true, "description": "Whether to generate synchronized audio for the video", "title": "Generate Audio", "type": "boolean", "x-sr-order": 501 }, "ratio": { "default": "adaptive", "description": "Aspect ratio of the video. 'adaptive' will auto-select based on prompt.", "enum": [ "21:9", "16:9", "4:3", "1:1", "3:4", "9:16", "adaptive" ], "title": "Ratio", "type": "string", "x-sr-order": 402 }, "resolution": { "default": "720p", "description": "Video resolution", "enum": [ "480p", "720p" ], "title": "Resolution", "type": "string", "x-sr-order": 401 } }, "required": [ "prompt" ], "title": "TextToVideoInput", "type": "object" } ``` ### Model output schema The model output schema is: ```json { "description": "Output schema for Seedance 2.0 video generation.", "properties": { "total_tokens": { "description": "Total tokens consumed for this generation", "title": "Total Tokens", "type": "integer" }, "units": { "description": "Billing units. 1000 units = $1.", "title": "Units", "type": "number" }, "video": { "properties": { "content_type": { "description": "The mime type of the file.", "title": "Content Type", "type": "string" }, "file_name": { "description": "The name of the file. It will be auto-generated if not provided.", "title": "File Name", "type": "string" }, "file_size": { "description": "The size of the file in bytes.", "title": "File Size", "type": "integer" }, "url": { "description": "The URL where the file can be downloaded from.", "title": "Url", "type": "string" } }, "required": [ "content_type", "file_name", "file_size", "url" ], "title": "SunraFile", "type": "object" } }, "required": [ "video", "total_tokens", "units" ], "title": "Seedance20Output", "type": "object" } ``` ## Example inputs and outputs Use the following example inputs and outputs to understand the model. ### Input example ```json { "prompt": "SUBJECTS:\nSubject 1: Lean kung fu chef with short, sharp-cut hair and defined jawline. Wears a modernized Chinese chef outfit fused with martial arts attire: fitted sleeveless upper garment with mandarin collar, dark matte fabric with subtle sheen; forearms wrapped with cloth bands for grip; loose tapered pants allowing wide stances; soft-soled shoes for silent footwork. Limbs slightly exaggerated in length; wrists highly flexible, elbows explosive. Movements follow clear martial rhythm (pause → burst → lock), with grounded footwork and fast pivots.\n\nENVIRONMENT:\nTraditional Chinese open-fire stove setup; iron wok fixed on stove; reflective metal counter; blue-and-white porcelain bowl placed in foreground center. Warm overhead light combined with intense dynamic firelight; light oil smoke and visible heat distortion.\n\nMOOD:\nHigh tension, aggressive, but controlled and precise; performance feels intentional and elegant rather than chaotic.\n\nTIMELINE:\n0:00-0:02: Extreme close-up, wide-angle POV. The chef stands on the ground behind the counter (counter always between camera and chef), steps in, compresses posture, then snaps his head up to lock eyes with camera. Both hands grip dough—he stretches it in one clean pull, folds once, then pulls again. The dough transitions from a single mass into a small number of even strands.\n\n0:02-0:05: Cut. Slight handheld motion. The chef keeps full control of the noodle bundle in both hands, stretching and aligning the strands into clean, parallel lines through fold → align → stretch cycles. Around 8–12 thin strands extend in smooth arcs, passing close to the lens in clean, readable motion. He finishes the pull and places the aligned noodle bundle neatly onto the counter in front of him.\n\n0:05-0:07: Move (slight recoil then stabilize). The chef lowers into a sliding step behind the counter. One palm sweeps across the surface—shrimp and greens lift and travel in a controlled arc into the wok. The wok remains on the stove; the chef grips only the handle. Any stray ingredient is clearly caught and redirected into the wok.\n\n0:07-0:10: Continuous shot. The chef controls the wok only through handle movement and qi force—no hands enter the wok. He lifts, tilts, and snaps the wok using wrist power; ingredients rise, rotate, and fall back into the wok. His elbow strikes the stove edge—sparks burst. He compresses the wok into the flame, then reverses to pull the flame upward into a second controlled flare.\n\n0:10-0:12: Match move tracking. The chef pivots while holding the wok handle. With the other hand, he picks up the same noodle bundle from the counter and feeds it directly into the wok in one controlled motion. He immediately resumes tossing using the wok—three clean toss cycles (fast → faster → stop). Shrimp, greens, and noodles integrate visibly; sauce coats evenly.\n\n0:12-0:15: Cut to stabilized POV. Sudden full stop. The chef tilts the wok and slides the finished dish into the porcelain bowl—visible: evenly coated noodles, plump shrimp, glossy greens. He lifts and presents the bowl directly toward the camera until it fills the frame. A finger taps the rim—subtle vibration. Steam rises rapidly, expanding until it fully covers the lens for a clean fade-out.\n", "duration": 15, "generate_audio": true, "ratio": "16:9", "resolution": "720p" } ``` ### Output example ```json { "total_tokens": 324900, "units": 2605.698, "video": { "content_type": "video/mp4", "file_name": "at7yafj5cvdmzgzabko6tl6spa.mp4", "file_size": 8541591, "url": "https://assets-cdn2.sunra.ai/resources/at7yafj5cvdmzgzabko6tl6spa.mp4" } } ``` ## Model code examples ### JavaScript ```javascript import { sunra } from "@sunra/client"; const result = await sunra.subscribe("bytedance/seedance-2.0/text-to-video", { input: { prompt: 'SUBJECTS:\nSubject 1: Lean kung fu chef with short, sharp-cut hair and defined jawline. Wears a modernized Chinese chef outfit fused with martial arts attire: fitted sleeveless upper garment with mandarin collar, dark matte fabric with subtle sheen; forearms wrapped with cloth bands for grip; loose tapered pants allowing wide stances; soft-soled shoes for silent footwork. Limbs slightly exaggerated in length; wrists highly flexible, elbows explosive. Movements follow clear martial rhythm (pause → burst → lock), with grounded footwork and fast pivots.\n\nENVIRONMENT:\nTraditional Chinese open-fire stove setup; iron wok fixed on stove; reflective metal counter; blue-and-white porcelain bowl placed in foreground center. Warm overhead light combined with intense dynamic firelight; light oil smoke and visible heat distortion.\n\nMOOD:\nHigh tension, aggressive, but controlled and precise; performance feels intentional and elegant rather than chaotic.\n\nTIMELINE:\n0:00-0:02: Extreme close-up, wide-angle POV. The chef stands on the ground behind the counter (counter always between camera and chef), steps in, compresses posture, then snaps his head up to lock eyes with camera. Both hands grip dough—he stretches it in one clean pull, folds once, then pulls again. The dough transitions from a single mass into a small number of even strands.\n\n0:02-0:05: Cut. Slight handheld motion. The chef keeps full control of the noodle bundle in both hands, stretching and aligning the strands into clean, parallel lines through fold → align → stretch cycles. Around 8–12 thin strands extend in smooth arcs, passing close to the lens in clean, readable motion. He finishes the pull and places the aligned noodle bundle neatly onto the counter in front of him.\n\n0:05-0:07: Move (slight recoil then stabilize). The chef lowers into a sliding step behind the counter. One palm sweeps across the surface—shrimp and greens lift and travel in a controlled arc into the wok. The wok remains on the stove; the chef grips only the handle. Any stray ingredient is clearly caught and redirected into the wok.\n\n0:07-0:10: Continuous shot. The chef controls the wok only through handle movement and qi force—no hands enter the wok. He lifts, tilts, and snaps the wok using wrist power; ingredients rise, rotate, and fall back into the wok. His elbow strikes the stove edge—sparks burst. He compresses the wok into the flame, then reverses to pull the flame upward into a second controlled flare.\n\n0:10-0:12: Match move tracking. The chef pivots while holding the wok handle. With the other hand, he picks up the same noodle bundle from the counter and feeds it directly into the wok in one controlled motion. He immediately resumes tossing using the wok—three clean toss cycles (fast → faster → stop). Shrimp, greens, and noodles integrate visibly; sauce coats evenly.\n\n0:12-0:15: Cut to stabilized POV. Sudden full stop. The chef tilts the wok and slides the finished dish into the porcelain bowl—visible: evenly coated noodles, plump shrimp, glossy greens. He lifts and presents the bowl directly toward the camera until it fills the frame. A finger taps the rim—subtle vibration. Steam rises rapidly, expanding until it fully covers the lens for a clean fade-out.\n', resolution: '720p', ratio: '16:9', duration: 15, generate_audio: true }, logs: true, onQueueUpdate: (update) => { console.log(`Status Update: ${update.status}, Request ID: ${update.request_id}`); }, }); console.log(result.data); console.log(result.requestId); ``` ### Python ```python import sunra_client result = sunra_client.subscribe( "bytedance/seedance-2.0/text-to-video", arguments={ "prompt": "SUBJECTS:\nSubject 1: Lean kung fu chef with short, sharp-cut hair and defined jawline. Wears a modernized Chinese chef outfit fused with martial arts attire: fitted sleeveless upper garment with mandarin collar, dark matte fabric with subtle sheen; forearms wrapped with cloth bands for grip; loose tapered pants allowing wide stances; soft-soled shoes for silent footwork. Limbs slightly exaggerated in length; wrists highly flexible, elbows explosive. Movements follow clear martial rhythm (pause → burst → lock), with grounded footwork and fast pivots.\n\nENVIRONMENT:\nTraditional Chinese open-fire stove setup; iron wok fixed on stove; reflective metal counter; blue-and-white porcelain bowl placed in foreground center. Warm overhead light combined with intense dynamic firelight; light oil smoke and visible heat distortion.\n\nMOOD:\nHigh tension, aggressive, but controlled and precise; performance feels intentional and elegant rather than chaotic.\n\nTIMELINE:\n0:00-0:02: Extreme close-up, wide-angle POV. The chef stands on the ground behind the counter (counter always between camera and chef), steps in, compresses posture, then snaps his head up to lock eyes with camera. Both hands grip dough—he stretches it in one clean pull, folds once, then pulls again. The dough transitions from a single mass into a small number of even strands.\n\n0:02-0:05: Cut. Slight handheld motion. The chef keeps full control of the noodle bundle in both hands, stretching and aligning the strands into clean, parallel lines through fold → align → stretch cycles. Around 8–12 thin strands extend in smooth arcs, passing close to the lens in clean, readable motion. He finishes the pull and places the aligned noodle bundle neatly onto the counter in front of him.\n\n0:05-0:07: Move (slight recoil then stabilize). The chef lowers into a sliding step behind the counter. One palm sweeps across the surface—shrimp and greens lift and travel in a controlled arc into the wok. The wok remains on the stove; the chef grips only the handle. Any stray ingredient is clearly caught and redirected into the wok.\n\n0:07-0:10: Continuous shot. The chef controls the wok only through handle movement and qi force—no hands enter the wok. He lifts, tilts, and snaps the wok using wrist power; ingredients rise, rotate, and fall back into the wok. His elbow strikes the stove edge—sparks burst. He compresses the wok into the flame, then reverses to pull the flame upward into a second controlled flare.\n\n0:10-0:12: Match move tracking. The chef pivots while holding the wok handle. With the other hand, he picks up the same noodle bundle from the counter and feeds it directly into the wok in one controlled motion. He immediately resumes tossing using the wok—three clean toss cycles (fast → faster → stop). Shrimp, greens, and noodles integrate visibly; sauce coats evenly.\n\n0:12-0:15: Cut to stabilized POV. Sudden full stop. The chef tilts the wok and slides the finished dish into the porcelain bowl—visible: evenly coated noodles, plump shrimp, glossy greens. He lifts and presents the bowl directly toward the camera until it fills the frame. A finger taps the rim—subtle vibration. Steam rises rapidly, expanding until it fully covers the lens for a clean fade-out.\n", "resolution": "720p", "ratio": "16:9", "duration": 15, "generate_audio": True }, with_logs=True, on_enqueue=print, on_queue_update=print, ) print(result) ``` ### Java ```java import ai.sunra.client.*; import java.util.Map; import com.google.gson.JsonObject; var client = SunraClient.withEnvCredentials(); var response = client.subscribe( "bytedance/seedance-2.0/text-to-video", SubscribeOptions.builder() .input(Map.of( "prompt", "SUBJECTS: Subject 1: Lean kung fu chef with short, sharp-cut hair and defined jawline. Wears a modernized Chinese chef outfit fused with martial arts attire: fitted sleeveless upper garment with mandarin collar, dark matte fabric with subtle sheen; forearms wrapped with cloth bands for grip; loose tapered pants allowing wide stances; soft-soled shoes for silent footwork. Limbs slightly exaggerated in length; wrists highly flexible, elbows explosive. Movements follow clear martial rhythm (pause → burst → lock), with grounded footwork and fast pivots. ENVIRONMENT: Traditional Chinese open-fire stove setup; iron wok fixed on stove; reflective metal counter; blue-and-white porcelain bowl placed in foreground center. Warm overhead light combined with intense dynamic firelight; light oil smoke and visible heat distortion. MOOD: High tension, aggressive, but controlled and precise; performance feels intentional and elegant rather than chaotic. TIMELINE: 0:00-0:02: Extreme close-up, wide-angle POV. The chef stands on the ground behind the counter (counter always between camera and chef), steps in, compresses posture, then snaps his head up to lock eyes with camera. Both hands grip dough—he stretches it in one clean pull, folds once, then pulls again. The dough transitions from a single mass into a small number of even strands. 0:02-0:05: Cut. Slight handheld motion. The chef keeps full control of the noodle bundle in both hands, stretching and aligning the strands into clean, parallel lines through fold → align → stretch cycles. Around 8–12 thin strands extend in smooth arcs, passing close to the lens in clean, readable motion. He finishes the pull and places the aligned noodle bundle neatly onto the counter in front of him. 0:05-0:07: Move (slight recoil then stabilize). The chef lowers into a sliding step behind the counter. One palm sweeps across the surface—shrimp and greens lift and travel in a controlled arc into the wok. The wok remains on the stove; the chef grips only the handle. Any stray ingredient is clearly caught and redirected into the wok. 0:07-0:10: Continuous shot. The chef controls the wok only through handle movement and qi force—no hands enter the wok. He lifts, tilts, and snaps the wok using wrist power; ingredients rise, rotate, and fall back into the wok. His elbow strikes the stove edge—sparks burst. He compresses the wok into the flame, then reverses to pull the flame upward into a second controlled flare. 0:10-0:12: Match move tracking. The chef pivots while holding the wok handle. With the other hand, he picks up the same noodle bundle from the counter and feeds it directly into the wok in one controlled motion. He immediately resumes tossing using the wok—three clean toss cycles (fast → faster → stop). Shrimp, greens, and noodles integrate visibly; sauce coats evenly. 0:12-0:15: Cut to stabilized POV. Sudden full stop. The chef tilts the wok and slides the finished dish into the porcelain bowl—visible: evenly coated noodles, plump shrimp, glossy greens. He lifts and presents the bowl directly toward the camera until it fills the frame. A finger taps the rim—subtle vibration. Steam rises rapidly, expanding until it fully covers the lens for a clean fade-out. ", "resolution", "720p", "ratio", "16:9", "duration", 15, "generate_audio", true)) .resultType(JsonObject.class) .onQueueUpdate(update -> System.out.printf( "\nStatus Update: %s, Request ID: %s%n", update.getStatus(), update.getRequestId() )) .logs(true) .build() ); System.out.println("Completed!"); System.out.println(response.getData()); ``` ### Kotlin ```kotlin import ai.sunra.client.kt.* import com.google.gson.JsonObject val client = createSunraClient() val response = client.subscribe( endpointId = "bytedance/seedance-2.0/text-to-video", input = mapOf( "prompt" to "SUBJECTS: Subject 1: Lean kung fu chef with short, sharp-cut hair and defined jawline. Wears a modernized Chinese chef outfit fused with martial arts attire: fitted sleeveless upper garment with mandarin collar, dark matte fabric with subtle sheen; forearms wrapped with cloth bands for grip; loose tapered pants allowing wide stances; soft-soled shoes for silent footwork. Limbs slightly exaggerated in length; wrists highly flexible, elbows explosive. Movements follow clear martial rhythm (pause → burst → lock), with grounded footwork and fast pivots. ENVIRONMENT: Traditional Chinese open-fire stove setup; iron wok fixed on stove; reflective metal counter; blue-and-white porcelain bowl placed in foreground center. Warm overhead light combined with intense dynamic firelight; light oil smoke and visible heat distortion. MOOD: High tension, aggressive, but controlled and precise; performance feels intentional and elegant rather than chaotic. TIMELINE: 0:00-0:02: Extreme close-up, wide-angle POV. The chef stands on the ground behind the counter (counter always between camera and chef), steps in, compresses posture, then snaps his head up to lock eyes with camera. Both hands grip dough—he stretches it in one clean pull, folds once, then pulls again. The dough transitions from a single mass into a small number of even strands. 0:02-0:05: Cut. Slight handheld motion. The chef keeps full control of the noodle bundle in both hands, stretching and aligning the strands into clean, parallel lines through fold → align → stretch cycles. Around 8–12 thin strands extend in smooth arcs, passing close to the lens in clean, readable motion. He finishes the pull and places the aligned noodle bundle neatly onto the counter in front of him. 0:05-0:07: Move (slight recoil then stabilize). The chef lowers into a sliding step behind the counter. One palm sweeps across the surface—shrimp and greens lift and travel in a controlled arc into the wok. The wok remains on the stove; the chef grips only the handle. Any stray ingredient is clearly caught and redirected into the wok. 0:07-0:10: Continuous shot. The chef controls the wok only through handle movement and qi force—no hands enter the wok. He lifts, tilts, and snaps the wok using wrist power; ingredients rise, rotate, and fall back into the wok. His elbow strikes the stove edge—sparks burst. He compresses the wok into the flame, then reverses to pull the flame upward into a second controlled flare. 0:10-0:12: Match move tracking. The chef pivots while holding the wok handle. With the other hand, he picks up the same noodle bundle from the counter and feeds it directly into the wok in one controlled motion. He immediately resumes tossing using the wok—three clean toss cycles (fast → faster → stop). Shrimp, greens, and noodles integrate visibly; sauce coats evenly. 0:12-0:15: Cut to stabilized POV. Sudden full stop. The chef tilts the wok and slides the finished dish into the porcelain bowl—visible: evenly coated noodles, plump shrimp, glossy greens. He lifts and presents the bowl directly toward the camera until it fills the frame. A finger taps the rim—subtle vibration. Steam rises rapidly, expanding until it fully covers the lens for a clean fade-out. ", "resolution" to "720p", "ratio" to "16:9", "duration" to 15, "generate_audio" to true), options = ai.sunra.client.kt.SubscribeOptions(logs = true), onUpdate = { update -> println("\nStatus Update: ${update.status}, Request ID: ${update.requestId}") } ) println("Completed!") println(response.data) ``` ### Curl ```bash curl --request POST \ --url https://api.sunra.ai/v1/queue/bytedance/seedance-2.0/text-to-video \ --header "Authorization: Key $SUNRA_KEY" \ --header "Content-Type: application/json" \ --data '{"prompt":"SUBJECTS:\nSubject 1: Lean kung fu chef with short, sharp-cut hair and defined jawline. Wears a modernized Chinese chef outfit fused with martial arts attire: fitted sleeveless upper garment with mandarin collar, dark matte fabric with subtle sheen; forearms wrapped with cloth bands for grip; loose tapered pants allowing wide stances; soft-soled shoes for silent footwork. Limbs slightly exaggerated in length; wrists highly flexible, elbows explosive. Movements follow clear martial rhythm (pause → burst → lock), with grounded footwork and fast pivots.\n\nENVIRONMENT:\nTraditional Chinese open-fire stove setup; iron wok fixed on stove; reflective metal counter; blue-and-white porcelain bowl placed in foreground center. Warm overhead light combined with intense dynamic firelight; light oil smoke and visible heat distortion.\n\nMOOD:\nHigh tension, aggressive, but controlled and precise; performance feels intentional and elegant rather than chaotic.\n\nTIMELINE:\n0:00-0:02: Extreme close-up, wide-angle POV. The chef stands on the ground behind the counter (counter always between camera and chef), steps in, compresses posture, then snaps his head up to lock eyes with camera. Both hands grip dough—he stretches it in one clean pull, folds once, then pulls again. The dough transitions from a single mass into a small number of even strands.\n\n0:02-0:05: Cut. Slight handheld motion. The chef keeps full control of the noodle bundle in both hands, stretching and aligning the strands into clean, parallel lines through fold → align → stretch cycles. Around 8–12 thin strands extend in smooth arcs, passing close to the lens in clean, readable motion. He finishes the pull and places the aligned noodle bundle neatly onto the counter in front of him.\n\n0:05-0:07: Move (slight recoil then stabilize). The chef lowers into a sliding step behind the counter. One palm sweeps across the surface—shrimp and greens lift and travel in a controlled arc into the wok. The wok remains on the stove; the chef grips only the handle. Any stray ingredient is clearly caught and redirected into the wok.\n\n0:07-0:10: Continuous shot. The chef controls the wok only through handle movement and qi force—no hands enter the wok. He lifts, tilts, and snaps the wok using wrist power; ingredients rise, rotate, and fall back into the wok. His elbow strikes the stove edge—sparks burst. He compresses the wok into the flame, then reverses to pull the flame upward into a second controlled flare.\n\n0:10-0:12: Match move tracking. The chef pivots while holding the wok handle. With the other hand, he picks up the same noodle bundle from the counter and feeds it directly into the wok in one controlled motion. He immediately resumes tossing using the wok—three clean toss cycles (fast → faster → stop). Shrimp, greens, and noodles integrate visibly; sauce coats evenly.\n\n0:12-0:15: Cut to stabilized POV. Sudden full stop. The chef tilts the wok and slides the finished dish into the porcelain bowl—visible: evenly coated noodles, plump shrimp, glossy greens. He lifts and presents the bowl directly toward the camera until it fills the frame. A finger taps the rim—subtle vibration. Steam rises rapidly, expanding until it fully covers the lens for a clean fade-out.\n","resolution":"720p","ratio":"16:9","duration":15,"generate_audio":true}' ``` ## Model readme >