video-prompt-generator（视频提示生成器）

Video Prompt Generator

Generate polished Sora 2 video prompts through the NanoPhoto.AI Prompt Generator API.

Prerequisites

1. Obtain an API key at: https://nanophoto.ai/settings/apikeys
2. Configure NANOPHOTO_API_KEY before using the skill.
3. Do not paste the API key into chat; store it in the platform's secure env setting for this skill.

Preferred OpenClaw setup:

• Open the skill settings for this skill
• Add an environment variable named NANOPHOTO_API_KEY
• Paste the API key as its value

Equivalent config shape:

{
  "skills": {
    "entries": {
      "video-prompt-generator": {
        "enabled": true,
        "env": {
          "NANOPHOTO_API_KEY": "your_api_key_here"
        }
      }
    }
  }
}

Other valid ways to provide the key:

• Shell: export NANOPHOTO_API_KEY="your_api_key_here"
• Tool-specific env config: any runtime that injects NANOPHOTO_API_KEY
• OpenClaw config fallback: the bundled script also falls back to ~/.openclaw/openclaw.json at skills.entries.video-prompt-generator.env.NANOPHOTO_API_KEY

Credential declaration summary:

• Primary credential: NANOPHOTO_API_KEY
• Resolution order in the bundled script: --api-key → NANOPHOTO_API_KEY environment variable → ~/.openclaw/openclaw.json skill env
• No unrelated credentials are required

Recommended workflow

1. Collect the user's topic or scene idea.
2. Decide whether the prompt is for textToVideo or imageToVideo.
3. Choose a technique, duration, locale, and optional model override.
4. For image-to-video mode, require one to three public image URLs.
5. Run the bundled script or call the API directly.
6. Return the generated prompt text exactly as produced unless the user asks for adaptation or rewriting.

Parameter guidance

• topic

- Required - Maximum 500 characters - Keep it specific enough to imply subject, mood, or motion

• mode

- textToVideo: default - imageToVideo: requires public imageUrls

• technique

- Default: montage - Choose the technique that best matches the intended visual language

• duration

- 10 - 15

• locale

- Default: en - Supported: en, zh, zh-TW, ja, ko, es, fr, de, pt, ru, ar

Preferred command

Use the bundled script for reliable prompt generation:

Text-to-video prompt

python3 scripts/video_prompt_generator.py \
  --topic "A serene Japanese garden with cherry blossoms falling into a koi pond" \
  --mode textToVideo \
  --technique slow-motion \
  --duration 15 \
  --locale en

Image-to-video prompt

python3 scripts/video_prompt_generator.py \
  --topic "Animate this landscape with gentle wind and floating clouds" \
  --mode imageToVideo \
  --technique long-take \
  --duration 10 \
  --locale en \
  --image-url https://example.com/landscape.jpg

The script resolves credentials in this order: --api-key, then NANOPHOTO_API_KEY from the environment, then ~/.openclaw/openclaw.json at skills.entries.video-prompt-generator.env.NANOPHOTO_API_KEY.

Output behavior

• The API returns streaming text, but the bundled script prints the final assembled prompt text
• Return the prompt directly when the user asked for prompt generation
• If the user wants editing or localization after generation, transform the generated prompt in a second step instead of changing API parameters retroactively

Manual API call

curl -X POST "https://nanophoto.ai/api/sora-2/generate-prompt" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $NANOPHOTO_API_KEY" \
  --data-raw '{
    "topic": "A futuristic cityscape at sunset with flying vehicles and neon lights",
    "technique": "aerial-view",
    "duration": 15,
    "locale": "en"
  }'

Error handling

Bundled files

• scripts/video_prompt_generator.py: generate a prompt from a topic using the NanoPhoto Prompt Generator API
• references/api.md: condensed API reference, inputs, and error behavior