How to Install and Use AI picture book generate in OpenClaw

Open your terminal and run the clawhub install command to add the ai-picture-book-generate skill to your OpenClaw instance. ClawHub downloads the skill package from the registry, validates it, and registers it with your local OpenClaw configuration. The installation takes under a minute on a standard internet connection. After installation completes, you will see a message listing the required configuration keys — typically an image generation API key and optionally a preferred model name. The skill will not activate until these are configured in the next steps. If you attempt to use it before configuring the image key, OpenClaw will return a clear error message pointing you to the missing configuration. If you see `clawhub: command not found`, your OpenClaw installation may be incomplete. Reinstall OpenClaw or verify that the ClawHub binary is on your system PATH. Run `clawhub --version` to confirm ClawHub is available before proceeding.

1clawhub install ai-picture-book-generate

The AI Picture Book Generator skill uses an image generation API to create illustrations for each page. The most commonly used providers are Stability AI (Stable Diffusion) and OpenAI (DALL-E 3). Each requires a separate API key obtained from their respective platforms. For Stability AI, visit platform.stability.ai, create an account, and generate an API key from the API Keys section. For OpenAI DALL-E, your existing OpenAI API key works — no separate key is needed if you already use OpenAI for other tasks. Set the key in your OpenClaw configuration using the clawhub config command or by editing your `.env` file directly. The exact environment variable name depends on which provider the skill version you installed targets — check with `clawhub info ai-picture-book-generate` to confirm the correct variable name. After setting the key, run `clawhub reload` to apply the configuration without restarting OpenClaw entirely.

1# For Stability AI (most common default)
2clawhub config set STABILITY_API_KEY sk-your-stability-key-here
3
4# For OpenAI DALL-E
5clawhub config set OPENAI_API_KEY sk-your-openai-key-here
6
7# Optionally set preferred illustration model
8clawhub config set PICTURE_BOOK_IMAGE_MODEL stable-diffusion-3
9
10# Verify config
11clawhub config get STABILITY_API_KEY

Open OpenClaw chat and type a story prompt. The more detail you provide — characters, setting, tone, age group, illustration style, and page count — the more precisely the skill tailors the output. For a first test, a simple prompt works fine to verify the skill is operational before investing time in a detailed creative brief. The skill processes your prompt in stages: first generating the story outline, then the page-level narrative text, then the illustration prompt for each page, then the actual illustrations, and finally assembling everything into the output document. For a 10-page book, this process typically takes 2-5 minutes depending on image generation API response times. While the book is generating, OpenClaw shows progress updates so you know which stage is in progress. If a single illustration fails (due to content policy or API timeout), the skill retries once automatically and skips with a placeholder if the retry also fails — the rest of the book is not affected. The output file is saved to your configured OpenClaw output directory. By default this is `~/openclaw-output/` but can be changed in your OpenClaw configuration.

The ai-picture-book-generate skill supports several configuration options that let you set defaults for all future books without repeating parameters in each prompt. The most commonly customized settings are illustration style, output format, page dimensions, and text font size. Illustration style presets include: `watercolor`, `cartoon`, `flat-design`, `pencil-sketch`, `photorealistic`, and `oil-painting`. Setting a default in your skill config means you only need to override it in the prompt when you want something different. Output format options include: `pdf` (single multi-page document), `png-pages` (individual PNG files per page), and `epub` if supported in your installed version. PDF is the most portable format for sharing; PNG pages are useful if you plan to import illustrations into another tool. RapidDev offers OpenClaw configuration reviews for teams building automated content pipelines — useful if you're generating large volumes of illustrated content and need to optimize for API cost and output consistency.

1# Customize skill defaults in ~/.openclaw/skills/ai-picture-book-generate.yaml
2ai-picture-book-generate:
3  default_style: watercolor        # watercolor | cartoon | flat-design | pencil-sketch
4  output_format: pdf               # pdf | png-pages | epub
5  page_width: 1024                 # pixels
6  page_height: 1024                # pixels (square for most picture books)
7  text_size: large                 # small | medium | large
8  min_pages: 5
9  max_pages: 20
10  illustrations_per_page: 1        # 1 full illustration per page

Once the picture book is generated and saved to your output directory, you can export it in multiple ways. The PDF format is ready to share immediately via email or cloud storage. The PNG pages format lets you import individual illustrations into tools like Canva, Figma, or Google Slides for further editing. To find your generated files, check the path reported by OpenClaw after generation completes. On most installations, generated books are saved to `~/openclaw-output/picture-books/` with a timestamp and the first few words of your title as the filename. You can also ask OpenClaw to regenerate specific pages if an illustration did not turn out as expected. Describe what you want changed and reference the page number — for example, 'Regenerate page 3 of the last picture book with a more detailed forest background and remove the character from the foreground.' The skill regenerates just that page and updates the assembled PDF. For teams sharing picture book output, configure your OpenClaw output directory to point to a shared folder or cloud sync location so colleagues can access completed books without manual file transfers.

1# Check your OpenClaw output directory
2ls ~/openclaw-output/picture-books/
3
4# Open the most recently generated PDF (macOS)
5open ~/openclaw-output/picture-books/$(ls -t ~/openclaw-output/picture-books/ | head -1)
6
7# Or configure a custom output directory
8clawhub config set OUTPUT_DIR /path/to/your/shared/folder