pixmind-video

Pixmind Video Generation Skill

Generate AI videos using Pixmind. Supports text-to-video and image-to-video generation.

Note: The API endpoint aihub-admin.aimix.pro is the official Pixmind API gateway. Result URLs on chatmix.top are Pixmind's CDN for generated content.

When to use

• User asks to generate or create a video
• User wants to animate an existing image into a video
• User requests video content from a text description

Prerequisites

1. Register at pixmind.io — 注册即送 200 积分免费试用
2. Create an API key in the API Platform dashboard
3. Set env PIXMIND_API_KEY with your key

API Details

Endpoint: POST https://aihub-admin.aimix.pro/api-platform/v1/generations Auth: Header Authorization: Bearer {API_KEY} (from env PIXMIND_API_KEY)

Request Body (JSON)

Available Models

Fetch the current model catalog from GET /api-platform/v1/models. The list below was verified against the API Platform generation service on 2026-08-10.

Unsupported Models (removed or never published in API)

The following slugs are rejected by the API with 不支持的模型. Do not use:

• All seedance-1.0* (1.0, 1.0-pro, 1.0-lite) — removed
• seedance-1.5, seedance-1.5-lite, seedance-1.5-turbo — only seedance-1.5-pro survives
• seedance-2.0, seedance-2.0-lite, seedance-2.0-mini, seedance-2.0-turbo, seedance-2.0-eco
• All kling-* (kling-v1.5, v1.6, std-1.6, v2, v2-master, v2-pro, v3, v3-motion-control, etc.) — Kling no longer exposed via this API
• All pixverse-* (1.0, 2.0, 3.0, 4.0, video, v5) — slugs partially recognized but route is broken (500 error: Cannot read properties of undefined (reading 'internalModel'))
• wan2.1-t2v, wan2.5-t2v, wan2.7-video, wan2.7-t2v, wan2.7-i2v
• All minimax-*, hailuo, hidream-*
• veo, veo-2, veo-3.2, veo-3.1-fast, veo-3.1-pro, veo-3.1-ultra, veo-3.0-pro, veo-3.0-ultra
• sora, sora-1, sora-3, sora-2.5, sora-2-fast, sora-2-mini, sora-2-ultra

Model Introductions

Seedance (ByteDance Doubao) — Flagship video generation series with smooth motion and realistic physics. Resolution parameter required.

• seedance-2.0-pro — Latest flagship. Best for professional-grade video.
• seedance-2.0-fast — Faster, lower cost variant of 2.0.
• seedance-1.5-pro — Previous flagship, still high quality.

Google Veo — Cinematic video generation with native audio support. Only 16:9 and 9:16 aspect ratios.

• veo-3.1 / veo-3.1-eco — Latest Veo 3.1 generation.
• veo-3.0 / veo-3 / veo-3.0-fast — Veo 3.0 generation.

OpenAI Sora — Strong at complex scenes and physical accuracy. Only durations 4, 8, or 12 seconds accepted.

• sora-2 — Latest stable Sora.
• sora-2-eco — Cheaper variant (provider availability varies).
• sora-2-pro — Pro tier (currently paused upstream).

Wan 2.6 (Alibaba) — Alibaba's video diffusion model. Supports all common aspect ratios.

• wan2.6-t2v — Text-to-video.
• wan2.6-i2v — Image-to-video.
• wan2.6-i2v-flash — Faster image-to-video variant.

Usage

Use curl or the included helper script:

# Text to video (via curl) — Seedance requires resolution
curl -X POST https://aihub-admin.aimix.pro/api-platform/v1/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $PIXMIND_API_KEY" \
  -d '{"type": "video", "prompt": "ocean waves", "model": "seedance-2.0-pro", "duration": 5, "resolution": "1080p", "aspectRatio": "16:9"}'

# Sora 2 — duration MUST be 4, 8, or 12
curl -X POST https://aihub-admin.aimix.pro/api-platform/v1/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $PIXMIND_API_KEY" \
  -d '{"type": "video", "prompt": "city night drive", "model": "sora-2", "duration": 8, "aspectRatio": "16:9"}'

# Veo 3.1 — only 16:9 or 9:16
curl -X POST https://aihub-admin.aimix.pro/api-platform/v1/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $PIXMIND_API_KEY" \
  -d '{"type": "video", "prompt": "timelapse", "model": "veo-3.1", "duration": 8, "aspectRatio": "16:9"}'

# Or use the helper script
node {baseDir}/video-generate.js --prompt "描述文字" --model seedance-2.0-pro --duration 5 --resolution 1080p

Task Status Polling

After generation, poll for results:

# Via curl
curl https://aihub-admin.aimix.pro/api-platform/v1/tasks/ \
  -H "Authorization: Bearer $PIXMIND_API_KEY"

# Or use the helper script
node {baseDir}/task-status.js --task-id  --poll

Response Format

Generate response:

{"code": 1000, "data": {"taskId": 19401, "status": "processing"}}

Task status response:

{
  "code": 1000,
  "data": {
    "taskId": 19401,
    "status": "ready",
    "progress": 100,
    "videoUrl": "https://chatmix.top/...",
    "coverUrl": "https://chatmix.top/..."
  }
}

• data.taskId — Use this to poll status
• Status values: processing → ready (success)
• On success: data.videoUrl contains the video URL, data.coverUrl has the cover image

Error Responses

• code: 400 with 不支持的模型: — Unsupported model slug.
• code: 500 with the parameter resolution specified in the request is not valid for model — Seedance models require an explicit resolution of 480p, 720p, or 1080p.
• code: 500 with Invalid aspect ratio: — Aspect ratio not supported by this model. Veo only accepts 16:9 / 9:16.
• code: 500 with invalid duration: s (model sora-2-preview only supports duration=[4 8 12]) — Sora only accepts durations of 4, 8, or 12 seconds.
• code: 1001 with 请输入提示词 — Empty prompt.
• code: 1001 with 速率限制：每minute最多60次请求 — Rate limited; back off.

Guidelines

1. Always confirm the prompt, duration, and model with the user before generating
2. Default to seedance-2.0-pro with resolution: 1080p and aspectRatio: 16:9 unless user specifies otherwise
3. Seedance models require resolution (480p / 720p / 1080p) — requests without it fail
4. Veo models only accept 16:9 or 9:16 aspect ratios
5. Sora 2 models only accept duration of 4, 8, or 12 seconds — pick one of these explicitly
6. If user provides a reference image, automatically use img2video mode and pick an image-to-video capable model (wan2.6-i2v, wan2.6-i2v-flash, or any Veo / Sora / Seedance model)
7. Video generation takes longer than images — use --poll with appropriate interval (recommend 5–10s)
8. After getting the task ID, poll until completion and return video URL
9. For faster results on a budget, prefer seedance-2.0-fast, wan2.6-i2v-flash, or veo-3.0-fast
10. Avoid probing the API by sending requests with arbitrary model names — many slugs that appear in the marketing site are not actually exposed by the API. Stick to the table above.