Blog to Kindle

Scrape blog/essay sites, compile into EPUB with cover art, and deliver to Kindle.

Quick Start

# 1. Fetch essays from a supported site
uv run scripts/fetch_blog.py --site paulgraham --output ./pg-essays

# 2. Generate cover (uses Nano Banana Pro)
# See nano-banana-pro skill for cover generation

# 3. Compile to EPUB with cover
uv run scripts/compile_epub.py --input ./pg-essays --cover ./cover.png --output essays.epub

# 4. Send to Kindle
uv run scripts/send_to_kindle.py --file essays.epub --kindle-email user@kindle.com

Workflow (MUST follow this order)

1. Fetch - Download all essays/posts from the blog
2. Generate Cover - Create cover art via Nano Banana Pro skill (DO NOT SKIP)
3. Compile - Combine into EPUB with cover embedded
4. Send - Email to Kindle address

⚠️ Always generate and include cover before sending. Never send without cover.

Supported Sites

Site	Key	URL Pattern
Paul Graham	paulgraham	paulgraham.com/articles.html
Kevin Kelly	kevinkelly	kk.org/thetechnium
Derek Sivers	sivers	sive.rs/blog
Wait But Why	waitbutwhy	waitbutwhy.com/archive
Astral Codex Ten	acx	astralcodexten.com

For unlisted sites, use --site custom --url <archive-url>.

Cover Generation

Use the nano-banana-pro skill to generate covers. Prompt template:

Book cover for '[Author Name]: [Subtitle]'. 
Minimalist design with elegant typography. 
[Brand color] accent. Clean white/cream background. 
Simple geometric or abstract motif related to [topic].
Professional literary feel. No photos, no faces.
Portrait orientation book cover dimensions.

Generate at 2K resolution for good quality without huge file size.

Kindle Delivery

Default Kindle address (Simon): simonpilkington74_8oVjpj@kindle.com

Uses Mail.app via AppleScript to send. Ensure:

• Sender email is on Kindle approved list
• File under 50MB (EPUB compresses well)

State Tracking

State files stored in ~/.clawdbot/state/blog-kindle/:

• {site}-last-fetch.json - Last fetch timestamp, article count
• {site}-sent.json - List of sent article IDs

Use for incremental updates (only fetch new posts).

Manual Workflow (no scripts)

If scripts unavailable, follow this pattern:

1. Fetch: curl archive page → parse article links → fetch each → convert to markdown
2. Combine: Concatenate markdown with YAML frontmatter (title, author)
3. Cover: Generate via Nano Banana Pro
4. Convert: pandoc combined.md -o output.epub --epub-cover-image=cover.png --toc
5. Send: AppleScript Mail.app with attachment

See references/manual-workflow.md for detailed steps.