text-to-video

Text To Video

Overview

Generate dynamic videos driven purely by text, covering 2 types of scenarios:

1. High-quality default mode: creative videos, advertising materials, copy visualization, concept demos; 4–15s, ratios adaptive/16:9/4:3/1:1/3:4/9:16/21:9, 480p/720p, audio on by default.
2. High-dynamics fast mode: pursuing dynamic effects / vivid visuals, or requiring a short 3s duration; 3–15s, only 16:9/1:1/9:16, 720p, audio off by default, stronger dynamic effects.

Not handled: image-driven video generation, audio-driven generation, video editing, adding effects/changing style to videos.

API Mapping

Scenario	Backend API
High-quality default mode	video_toffee_t2v_v20
High-dynamics fast mode	video_bonbon_txt2vid_v30

model mapping:

• auto(default) / not passed → decided by routing rules
• toffee → video_toffee_t2v_v20
• bonbon → video_bonbon_txt2vid_v30
• Mutual fallback: if one fails, switch to the other and retry

Dependencies

• meitu-cli: >=2.0.6
• Credentials：CONFIG AKSK → meitu tools update; EXEC AKSK → run command (see root CONFIG.md）
• Environment variables：MEITU_OPENAPI_TOOL_TASK_MODE=command

Core Workflow

Preflight → Execute → Deliver

Preflight

1. meitu --version ≥ 2.0.6
2. meitu auth verify --json
3. Confirm registry contains text-to-video
4. Resolve output_dir:openclaw.yaml → ./output/；else → ~/.openclaw/workspace/visual/output/text-to-video/
5. Duration clamping: user requests <3s → clamp to 3s and inform

Execute

Trigger signals / routing rules

Decision order:

1. Duration <3s → clamp to 3s
2. Dynamics priority / 3s short duration matched → candidate video_bonbon_txt2vid_v30
3. If ratio is 4:3/3:4/21:9/adaptive → ratio priority, switch back to video_toffee_t2v_v20
4. No special requirements → video_toffee_t2v_v20(default)

Scenario	Decision keywords	Routing
Dynamics priority	dynamic, vivid	video_bonbon_txt2vid_v30
3s short duration	3 seconds	video_bonbon_txt2vid_v30
Default / no special requirements	General	video_toffee_t2v_v20
Special ratios (4:3/3:4/21:9/adaptive)	--	Force video_toffee_t2v_v20(ratio priority)
Want 480p	--	video_toffee_t2v_v20(bonbon does not support 480p)

Parameter definitions

General parameters:

Parameter	Type	Required	Range	Default value	Description
prompt	STRING	Yes	--	--	Video content description. If missing → ask follow-up
video_duration	NUMBER	Yes	3–15	Varies by API	toffee: 4–15, default 10; bonbon: 3–15, default 5
sound	STRING / BOOL	Yes	on/off or true/false	Varies by API	toffee: BOOL generate_audio, default true; bonbon: STRING on/off, default off
aspect_ratio	STRING	Yes	adaptive / 16:9 / 1:1 / 9:16 / 4:3 / 3:4 / 21:9	Varies by API	toffee supports all ratios, default adaptive; bonbon only 16:9/1:1/9:16, default 16:9
resolution	STRING	No	480p / 720p	720p	Only supported by toffee; bonbon is fixed at 720p

Parameter mapping for each API:

video_toffee_t2v_v20(high-quality default)

Unified parameter	API parameter	Mapping rule
prompt	prompt	Direct pass-through
video_duration	video_duration	clamp(4, 15), default 10
sound	generate_audio	on→true / off→false; default true
aspect_ratio	ratio	Full-ratio direct pass-through, default adaptive
resolution	resolution	480p/720p, default 720p

video_bonbon_txt2vid_v30(high dynamic fast)

Unified parameters	API parameters	Mapping rules
prompt	prompt	Direct pass-through
video_duration	video_duration	clamp(3, 15), default 5
sound	sound	on/off direct pass-through, default off
aspect_ratio	aspect_ratio	Supports 16:9/1:1/9:16; other ratios switch to toffee

Tool invocation

Default high quality:

meitu text-to-video \
  --skill_name skill_text-to-video \
  --prompt "{video_description}" \
  --video_duration 10 \
  --aspect_ratio adaptive \
  --resolution 720p \
  --sound true \
  --json --download-dir {output_dir}

High dynamic fast (3s):

meitu text-to-video \
  --skill_name skill_text-to-video \
  --prompt "{dynamic_scene}" \
  --model bonbon \
  --video_duration 3 \
  --aspect_ratio 16:9 \
  --sound off \
  --json --download-dir {output_dir}

Error fallback

Scenario	Handling method
prompt missing	Prompt "Please describe the video content you want to generate"
Duration <3s	Clamp to 3s, notify "Text-to-video supports a minimum of 3 seconds"
Duration <4s + default mode	toffee minimum 4s → fallback video_bonbon_txt2vid_v30
Special ratio + fast mode	bonbon only 16:9/1:1/9:16 → switch video_toffee_t2v_v20, notify "The current ratio does not support dynamic mode, switched to default mode"
Duration unspecified + default mode	Default 10s
Duration unspecified + fast mode	Default 5s
video_toffee_t2v_v20 failed	switch video_bonbon_txt2vid_v30 retry
video_bonbon_txt2vid_v30 failed	switch video_toffee_t2v_v20 retry → still fails, return error
Timeout >600s	Reduce duration / resolution and retry once → still fails, return error
Content compliance interception	Directly return compliance prompt, no retry, no fallback

Conflict rules:

• Dynamic + special ratio → ratio takes priority, use video_toffee_t2v_v20
• 3s duration + default mode → duration takes priority, use video_bonbon_txt2vid_v30
• Want 480p → use video_toffee_t2v_v20(bonbon does not support)

Deliver

Parse --json：

• ok: true → downloaded_files[0].saved_path as local video
• ok: false → output code + hint

Persist to disk:mv {file} {output_dir}/{YYYY-MM-DD}_{descriptive-name}.mp4

Output

• Format: MP4
• Naming: {YYYY-MM-DD}_{descriptive-name}.mp4
• Location: project → ./output/; one-time → ~/.openclaw/workspace/visual/output/text-to-video/

Baseline Task ID

See root directory references/task-id-baseline.md in text-to-video entry.