seedance-2.5

ByteDance's next-generation video model: up to 30 seconds per clip with native 4K, substantially better instruction following and long-form narrative. Supports multi-modal references (image + video + audio) and first/last frame control.

Overview

Field	Value
Model ID	`doubao-seedance-2-5`
CLI	`dlazy seedance-2.5`
MCP	`seedance-2.5` (Claude Code surfaces this as `mcp__dlazy__seedance-2.5`)
Type	`video`
Execution	Async task; the CLI polls until completion (`--no-wait` returns `generateId` immediately)
Batch	Supports `--batch <n>` parallel fan-out

Parameters

Arg	Type	Required	Notes
`prompt`	string	No	Prompt
`generation_mode`	"components" \| "frames"	No	Generation Mode（components=Components; frames=Frames）; default `"components"`
`images`	array<url>	No	Images; image (accepts URL, local path, or data: URL); max 20 items; only when !(generation_mode="frames")
`videos`	array<url>	No	Videos; video (accepts URL, local path, or data: URL); max 5 items; only when !(generation_mode="frames")
`audios`	array<url>	No	Audios; audio (accepts URL, local path, or data: URL); max 5 items; only when !(generation_mode="frames")
`firstFrame`	url	No	First Frame; image (accepts URL, local path, or data: URL); only when generation_mode="frames"
`lastFrame`	url	No	Last Frame; image (accepts URL, local path, or data: URL); only when generation_mode="frames"
`resolution`	"480p" \| "720p" \| "1080p" \| "2160p"	No	Resolution; default `"720p"`
`ratio`	"21:9" \| "16:9" \| "4:3" \| "1:1" \| "3:4" \| "9:16"	No	Aspect Ratio; default `"16:9"`
`duration`	"4" \| … \| "15" \| "20" \| "25" \| "30"	No	Duration (s); default `"5"`
`generate_audio`	string	No	Generate Audio; default `true`

--input @file.json or --input '{...}' can supply all args at once; flags take precedence over --input keys.

CLI Examples

dlazy seedance-2.5 --help
dlazy seedance-2.5 --prompt "Write your prompt here" --generation_mode "components" --images "https://example.com/reference1.jpg"
dlazy seedance-2.5 --prompt "Write your prompt here" --generation_mode "components" --images "https://example.com/reference1.jpg" --dry-run
dlazy seedance-2.5 --prompt "Write your prompt here" --generation_mode "components" --images "https://example.com/reference1.jpg" --no-wait
dlazy seedance-2.5 --prompt "Write your prompt here" --generation_mode "components" --images "https://example.com/reference1.jpg" --batch 4

Compose with a pipeline

dlazy gpt-image-2 --prompt "reference visual" \
  | dlazy seedance-2.5 --images - --prompt "Write your prompt here"

MCP

MCP is consumed by AI clients, not handwritten. Once dLazy is registered as an MCP server the tool appears in the client's tool list — Claude Code surfaces it as mcp__dlazy__seedance-2.5; generic clients (e.g. OpenClaw) call it as seedance-2.5.

See MCP setup for how to add the server.

Output

{
  "outputs": [
    { "type": "video", "id": "o_...", "url": "https://files.dlazy.com/result.mp4", "mimeType": "video/mp4" }
  ]
}

Media tools emit image / video / audio / file outputs. Use --output url to print only the URLs on stdout.