Wan 2.7 Text-to-Video Spicy

Wan 2.7 Text-to-Video Spicy API

Overview

Wan 2.7 Text-to-Video Spicy is the spicy route to Wan 2.7’s audio-enabled, multi-shot video model. It turns text prompts into 4–15 second videos at 720P or 1080P and can interpret shot changes directly from natural-language instructions.

Users can describe a single continuous shot, divide a sequence with labels such as Shot 1 and Shot 2, or assign timestamps such as Shot 1 [0–3s]. When no shot structure is provided, Wan 2.7 can interpret the prompt and decide whether the scene should remain continuous or move through multiple shots.

The model is designed to keep the main subject coherent across transitions while generating synchronized native audio. The iCreat endpoint exposes Prompt, Negative Prompt, resolution, aspect ratio, duration, and watermark controls.

SP means spicy: an uncensored version designed for mature NSFW creation without standard content filters. Access is limited to adults aged 18 and over and remains subject to applicable laws and prohibited-content rules.

Use the model through:

aliyun/wan2-7/text-to-video-sp

Key Features

Natural-language multi-shot direction. Wan 2.7 does not require a dedicated Shot Type parameter. Describe the sequence directly in the Prompt with shot labels, timestamps, camera positions, and actions.

Subject continuity across transitions. The model is designed to maintain the main subject as the video moves between wide shots, close-ups, and other scene changes. Clear descriptions of identity, clothing, environment, and movement improve continuity.

Synchronized native audio. Wan 2.7 generates native audio with the video automatically, allowing the visual sequence and sound to develop in the same generation. No audio upload or separate audio request is required.

Flexible 4–15 second storytelling. Choose any integer duration from 4 to 15 seconds. Short clips can focus on one action, while longer outputs can use timestamps to organize several connected visual beats.

Negative Prompt control. Use negative_prompt to reduce unwanted content, visual artifacts, incorrect anatomy, style changes, cuts, or camera behavior. It can improve control but does not guarantee complete removal.

Spicy Text-to-Video access. Wan 2.7 SP removes standard content filters for lawful mature 18+ generation while preserving the model’s multi-shot and audio-video capabilities.

Model Comparison

Wan 2.7 SP vs Wan 2.7

Wan 2.7 SP vs Wan 2.6 T2V vs HappyHorse 1.1 Spicy vs Vidu Q3-Pro T2V

Inputs

Parameters

Pricing

Total Cost = Unit Price × Output Video Duration

Quick Start

curl --fail-with-body --connect-timeout 10 --max-time 60 \
  -X POST https://api.icreat.ai/v1/task/submit/aliyun/wan2-7/text-to-video-sp \
  -H "Authorization: Bearer ${ICREAT_API_KEY}" \
  -H "Content-Type: application/json" \
  -d '{
    "input": {
      "prompt": "A cinematic nighttime scene in a private penthouse, warm practical lighting, slow camera push-in, realistic movement, detailed environment",
      "negative_prompt": "distorted anatomy, extra limbs, unstable face, flickering, sudden cut, abrupt camera movement, text, logo"
    },
    "parameters": {
      "resolution": "1080P",
      "ratio": "16:9",
      "duration": 8,
      "watermark": false
    }
  }'

The request returns a task_id.

Submit request → Receive task_id → Poll status → Retrieve result

Query the task status until it reaches SUCCEEDED, then use the same task_id to retrieve the result.

Output

Wan 2.7 Text-to-Video SP uses an asynchronous task flow.

The submission request returns a task_id. Task status is checked through the query-status endpoint, and the completed result is retrieved through the get-result endpoint after the task succeeds.

Detailed result fields are not listed until they are confirmed from an actual iCreat response.

Use Cases

Multi-shot narrative sequences. Build short stories with explicit shot labels or timestamps while keeping the main character and scene direction connected across transitions.

Cinematic montage concepts. Move between establishing shots, medium shots, and close-ups for trailers, mood pieces, music-video concepts, or visual storyboards.

Mature fictional storytelling. Generate lawful adult-oriented scenes that need more narrative development than a single continuous action.

Cinematic previsualization. Test scene order, camera position, pacing, lighting, and action before committing to a longer production workflow.

Synchronized audio-video concepts. Generate visual sequences and native sound together for concepts where atmosphere, transitions, and pacing need to develop as one piece.

Limitations

Wan 2.7 Text-to-Video SP is available only to adults aged 18 and over. Requests remain subject to applicable laws and prohibited-content rules. Removing standard content filters does not permit illegal content.

The endpoint is pure Text-to-Video and does not accept image, video, or audio input. Output length is limited to 4–15 seconds.

Wan 2.7 supports multi-shot direction through natural-language prompts, but the endpoint does not expose separate Shot Type, per-shot duration, or per-shot parameter fields. Shot timing must be written inside the Prompt, and the model may interpret transitions differently from the requested structure.

Subject identity, anatomy, motion, continuity, and audio timing can still vary across shots. More scene changes create more opportunities for visual drift. Negative Prompts can reduce unwanted results but cannot guarantee complete removal.

FAQ

What does SP mean in Wan 2.7?

SP means spicy. It is the uncensored version of Wan 2.7 Text-to-Video, designed for mature NSFW generation without standard content filters. Access is limited to adults aged 18 and over and remains subject to applicable laws and prohibited-content rules.

Is Wan 2.7 SP completely unrestricted?

No. SP removes standard content filtering, but illegal content and other prohibited uses are not allowed. It should not be described as unlimited, rule-free, or open to every request.

How should a multi-shot Prompt be structured?

Start with a short description of the complete sequence, then divide it into labeled shots. For example: Shot 1 [0–3s]: wide establishing shot; Shot 2 [3–6s]: medium shot of the subject moving; Shot 3 [6–8s]: close-up reaction. Keep the subject, environment, and direction of movement consistent between shots.

Does Wan 2.7 decide between single-shot and multi-shot automatically?

Yes. When the Prompt does not define a shot structure, the model can interpret the scene and decide whether to keep one continuous shot or introduce transitions. State single-shot video or provide labeled shots when a specific structure matters.

Does Wan 2.7 SP generate native audio automatically?

Yes. Native audio is generated with the video automatically. No audio input or separate audio-generation request is required.

Can Wan 2.7 SP generate silent video?

The current endpoint does not expose an audio switch. A separate post-production step is required when a silent final file is needed.

How should a Negative Prompt be written?

List the specific content, visual errors, styles, cuts, or camera behavior you want to avoid. Short phrases such as extra limbs, flickering, unstable face, abrupt cut, or inconsistent clothing are clearer than long contradictory instructions.

When should 1080P be used instead of 720P?

Use 720P for lower-cost testing and repeated Prompt iteration. Use 1080P for selected outputs that need more visible detail or a larger delivery size.

Can it use images, videos, or audio files?

No. This endpoint is pure Text-to-Video and does not accept image, video, or audio input.