feat: add ByteDance Seedance video generation models (seedance-2.0, seedance-2.0-fast, seedance-1.5-pro)

apps/gateway/src/videos/videos.ts

@@ -911,6 +911,15 @@ function getAvalancheSoraTaskModelName(
 	return `${baseModelName}-${inputMode === "reference" ? "image" : "text"}-to-video`;
 }
 
+function getBytedanceVideoUpstreamModelName(baseModelName: string): string {
+	const modelNameMap: Record<string, string> = {
+		"seedance-2-0": "dreamina-seedance-2-0-260128",
+		"seedance-2-0-fast": "dreamina-seedance-2-0-fast-260128",
+		"seedance-1-5-pro": "seedance-1-5-pro-251215",
+	};
+	return modelNameMap[baseModelName] ?? baseModelName;
+}
+
 function getVideoUpstreamModelName(
 	providerId: Provider,
 	baseModelName: string,
@@ -920,6 +929,8 @@ function getVideoUpstreamModelName(
 	switch (providerId) {
 		case "avalanche":
 			return getAvalancheVideoModelName(baseModelName);
+		case "bytedance":
+			return getBytedanceVideoUpstreamModelName(baseModelName);
 		case "google-vertex":
 			return baseModelName;
 		default:
@@ -969,6 +980,8 @@ function getDefaultVideoProviderBaseUrl(providerId: Provider): string | null {
 	switch (providerId) {
 		case "openai":
 			return "https://api.openai.com";
+		case "bytedance":
+			return "https://ark.ap-southeast.bytepluses.com/api/v3";
 		case "google-vertex":
 			return "https://aiplatform.googleapis.com";
 		default:
@@ -2804,6 +2817,108 @@ async function createGoogleVertexVideoJob(
 	};
 }
 
+function getBytedanceVideoAspectRatio(videoSize: VideoSizeConfig): string {
+	if (videoSize.orientation === "portrait") {
+		return "9:16";
+	}
+	return "16:9";
+}
+
+async function createBytedanceVideoJob(
+	providerContext: ProviderContext,
+	providerMapping: ProviderModelMapping,
+	videoSize: VideoSizeConfig,
+	prompt: string,
+	durationSeconds: number,
+	includeAudio: boolean,
+	firstFrameInput: VideoImageInput | undefined,
+	processedFirstFrame: ProcessedVideoImageInput | null,
+	processedReferenceImages: ProcessedVideoImageInput[],
+): Promise<{
+	upstreamId: string;
+	upstreamRequest: Record<string, unknown>;
+	upstreamResponse: Record<string, unknown>;
+}> {
+	const upstreamModelName = getBytedanceVideoUpstreamModelName(
+		providerMapping.modelName,
+	);
+	const content: Array<Record<string, unknown>> = [
+		{
+			type: "text",
+			text: prompt,
+		},
+	];
+
+	if (processedFirstFrame) {
+		content.push({
+			type: "image_url",
+			image_url: {
+				url: `data:${processedFirstFrame.mimeType};base64,${processedFirstFrame.bytesBase64Encoded}`,
+			},
+		});
+	}
+
+	if (processedReferenceImages.length > 0) {
+		for (const image of processedReferenceImages) {
+			content.push({
+				type: "image_url",
+				image_url: {
+					url: `data:${image.mimeType};base64,${image.bytesBase64Encoded}`,
+				},
+			});
+		}
+	}
+
+	const isDreaminaModel = upstreamModelName.startsWith("dreamina-");
+
+	const upstreamRequest: Record<string, unknown> = {
+		model: upstreamModelName,
+		content,
+		duration: isDreaminaModel ? `${durationSeconds}s` : durationSeconds,
+		aspect_ratio: getBytedanceVideoAspectRatio(videoSize),
+		generate_audio: includeAudio,
+	};
+
+	if (isDreaminaModel) {
+		upstreamRequest.resolution =
+			videoSize.resolution === "1080p" ? "1080p" : "720p";
+	}
+
+	const upstreamUrl = joinUrl(
+		providerContext.baseUrl,
+		"/contents/generations/tasks",
+	);
+	const rawResponse = await fetchUpstreamJson(upstreamUrl, {
+		method: "POST",
+		headers: {
+			"Content-Type": "application/json",
+			...getProviderHeaders("bytedance", providerContext.token, {
+				requestId: providerContext.requestId,
+			}),
+		},
+		body: JSON.stringify(upstreamRequest),
+	});
+
+	const upstreamResponse = addRequestedVideoMetadata(
+		{
+			...rawResponse,
+			status: rawResponse.status ?? "queued",
+			duration: durationSeconds,
+			aspect_ratio: upstreamRequest.aspect_ratio,
+		},
+		videoSize,
+	);
+
+	const upstreamId = extractUpstreamVideoId(upstreamResponse);
+	if (!upstreamId) {
+		throw new HTTPException(502, {
+			message: "ByteDance video response did not include a task id",
+		});
+	}
+
+	return { upstreamId, upstreamRequest, upstreamResponse };
+}
+
 async function createUpstreamVideoJob(
 	providerContext: ProviderContext,
 	providerMapping: ProviderModelMapping,
@@ -2856,6 +2971,18 @@ async function createUpstreamVideoJob(
 						lastFrameInput,
 						referenceImageInputs,
 					);
+		case "bytedance":
+			return await createBytedanceVideoJob(
+				providerContext,
+				providerMapping,
+				videoSize,
+				prompt,
+				durationSeconds,
+				includeAudio,
+				firstFrameInput,
+				processedFirstFrame,
+				processedReferenceImages,
+			);
 		case "google-vertex":
 			return await createGoogleVertexVideoJob(
 				providerContext,

apps/worker/src/services/video-jobs.ts

@@ -136,6 +136,8 @@ function getDefaultVideoProviderBaseUrl(providerId: Provider): string | null {
 	switch (providerId) {
 		case "openai":
 			return "https://api.openai.com";
+		case "bytedance":
+			return "https://ark.ap-southeast.bytepluses.com/api/v3";
 		case "google-vertex":
 			return "https://aiplatform.googleapis.com";
 		default:
@@ -1791,6 +1793,72 @@ async function finalizeVideoJob(job: VideoJobRecord): Promise<void> {
 	}
 }
 
+function isBytedanceVideoProvider(providerId: string): boolean {
+	return providerId === "bytedance";
+}
+
+async function fetchBytedanceStatus(
+	job: VideoJobRecord,
+	providerContext: ResolvedVideoProviderContext,
+): Promise<Record<string, unknown>> {
+	const url = joinUrl(
+		providerContext.baseUrl,
+		`/contents/generations/tasks/${job.upstreamId}`,
+	);
+	const { body, response } = await fetchJsonResponse(url, {
+		method: "GET",
+		headers: getVideoProviderHeaders(job, providerContext),
+	});
+
+	if (!response.ok) {
+		throw new Error(
+			typeof body.error === "object" &&
+			body.error &&
+			"message" in body.error &&
+			typeof body.error.message === "string"
+				? body.error.message
+				: `ByteDance status request failed with status ${response.status}`,
+		);
+	}
+
+	const data =
+		body.data && typeof body.data === "object"
+			? (body.data as Record<string, unknown>)
+			: body;
+
+	const rawStatus = typeof data.status === "string" ? data.status : "queued";
+	const content =
+		data.content && typeof data.content === "object"
+			? (data.content as Record<string, unknown>)
+			: null;
+	const videoUrl =
+		content && typeof content.video_url === "string" ? content.video_url : null;
+
+	return addRequestedVideoMetadata(job, {
+		...body,
+		status: rawStatus,
+		progress:
+			rawStatus === "succeeded" || rawStatus === "completed"
+				? 100
+				: rawStatus === "failed"
+					? 100
+					: rawStatus === "running"
+						? 50
+						: 0,
+		url: videoUrl,
+		video_url: videoUrl,
+		output_url: videoUrl,
+		mime_type: videoUrl ? "video/mp4" : undefined,
+		error:
+			rawStatus === "failed"
+				? (extractError(data) ?? {
+						message: "ByteDance video generation failed",
+					})
+				: null,
+		bytedance_raw_response: body,
+	});
+}
+
 async function fetchUpstreamStatus(
 	job: VideoJobRecord,
 ): Promise<Record<string, unknown>> {
@@ -1806,6 +1874,10 @@ async function fetchUpstreamStatus(
 		return await fetchGoogleVertexStatus(job, providerContext);
 	}
 
+	if (isBytedanceVideoProvider(job.usedProvider)) {
+		return await fetchBytedanceStatus(job, providerContext);
+	}
+
 	return await fetchGenericVideoStatus(job, providerContext);
 }
 

packages/models/src/models/bytedance.ts

@@ -128,6 +128,114 @@ export const bytedanceModels = [
 			},
 		],
 	},
+	{
+		id: "seedance-2-0",
+		name: "Seedance 2.0",
+		description:
+			"ByteDance Seedance 2.0 high-quality video generation model with text-to-video and image-to-video support",
+		family: "bytedance",
+		output: ["video"],
+		releasedAt: new Date("2025-12-18"),
+		providers: [
+			{
+				test: "skip",
+				providerId: "bytedance",
+				modelName: "seedance-2-0",
+				inputPrice: undefined,
+				outputPrice: undefined,
+				requestPrice: undefined,
+				perSecondPrice: {
+					default_audio: 0.1512,
+					default_video: 0.1512,
+					"1080p_audio": 0.3402,
+					"1080p_video": 0.3402,
+				},
+				contextSize: 2000,
+				maxOutput: 1,
+				streaming: false,
+				vision: false,
+				tools: false,
+				jsonOutput: false,
+				videoGenerations: true,
+				supportedVideoSizes: ["1280x720", "720x1280", "1920x1080", "1080x1920"],
+				supportedVideoDurationsSeconds: [5, 10],
+				supportsVideoAudio: true,
+				supportsVideoWithoutAudio: true,
+			},
+		],
+	},
+	{
+		id: "seedance-2-0-fast",
+		name: "Seedance 2.0 Fast",
+		description:
+			"ByteDance Seedance 2.0 Fast - accelerated video generation model for quicker outputs",
+		family: "bytedance",
+		output: ["video"],
+		releasedAt: new Date("2025-12-18"),
+		providers: [
+			{
+				test: "skip",
+				providerId: "bytedance",
+				modelName: "seedance-2-0-fast",
+				inputPrice: undefined,
+				outputPrice: undefined,
+				requestPrice: undefined,
+				perSecondPrice: {
+					default_audio: 0.05184,
+					default_video: 0.02592,
+					"1080p_audio": 0.1166,
+					"1080p_video": 0.05832,
+				},
+				contextSize: 2000,
+				maxOutput: 1,
+				streaming: false,
+				vision: false,
+				tools: false,
+				jsonOutput: false,
+				videoGenerations: true,
+				supportedVideoSizes: ["1280x720", "720x1280", "1920x1080", "1080x1920"],
+				supportedVideoDurationsSeconds: [5, 10],
+				supportsVideoAudio: true,
+				supportsVideoWithoutAudio: true,
+			},
+		],
+	},
+	{
+		id: "seedance-1-5-pro",
+		name: "Seedance 1.5 Pro",
+		description:
+			"ByteDance Seedance 1.5 Pro - professional video generation model with complex instruction following and native audio",
+		family: "bytedance",
+		output: ["video"],
+		releasedAt: new Date("2025-09-15"),
+		providers: [
+			{
+				test: "skip",
+				providerId: "bytedance",
+				modelName: "seedance-1-5-pro",
+				inputPrice: undefined,
+				outputPrice: undefined,
+				requestPrice: undefined,
+				perSecondPrice: {
+					default_audio: 0.05184,
+					default_video: 0.02592,
+					"1080p_audio": 0.1166,
+					"1080p_video": 0.05832,
+				},
+				contextSize: 2000,
+				maxOutput: 1,
+				streaming: false,
+				vision: false,
+				tools: false,
+				jsonOutput: false,
+				videoGenerations: true,
+				supportedVideoSizes: ["1280x720", "720x1280", "1920x1080", "1080x1920"],
+				supportedVideoDurationsSeconds: [5, 10],
+				supportsVideoAudio: true,
+				supportsVideoWithoutAudio: true,
+			},
+		],
+	},
 	{
 		id: "seedream-4-5",
 		name: "Seedream 4.5",