Agent-Skills.md

Agent Skills: WeRead Text Downloader

Use when downloading purchased WeRead books into chapter-based plain text for AI reading, using a CLI workflow that logs into the WeRead web reader and writes metadata, table of contents, and chapter txt files.

UncategorizedID: flying3615/my-skills-writer/weread-text-downloader

Repository

flying3615/my-skills-writer
flying3615

Install this agent skill to your local

pnpm dlx add-skill https://github.com/flying3615/my-skills-writer/tree/HEAD/weread-text-downloader

Skill Files

Browse the full folder contents for weread-text-downloader.

Download Skill

Loading file tree…

weread-text-downloader/SKILL.md

Skill Metadata

Name
weread-text-downloader
Description
Use when downloading purchased WeRead books into chapter-based plain text for AI reading, using a CLI workflow that logs into the WeRead web reader and writes metadata, table of contents, and chapter txt files.

WeRead Text Downloader

Use this skill to download a purchased WeRead book and export it as chapter-based plain text.

Default behavior:

  • CLI-only workflow
  • final output is plain text, not epub
  • chapter-based output for downstream AI reading
  • no intermediate rdata.zip is kept

Dependencies

Required:

  • python3
  • pyppeteer

Optional for --headless QR display:

  • pillow
  • pyzbar
  • qrcode

Install the minimum dependency set with:

pip install pyppeteer

Install the headless extras only if needed:

pip install pillow pyzbar qrcode

Command

python3 weread-text-downloader/scripts/weread_text.py download "Book Name"

Optional flags:

  • --out-dir ./out
  • --headless
  • --delay 2
  • --verbose

Example:

python3 weread-text-downloader/scripts/weread_text.py download "怦然心动" \
  --out-dir ./out \
  --verbose

Output

The script writes:

out/
  <book-slug>/
    metadata.json
    toc.json
    chapters/
      001-<chapter-slug>.txt
      002-<chapter-slug>.txt

Use metadata.json and toc.json for structure-aware downstream processing. Use chapters/*.txt as the primary AI reading input.

Workflow

  1. Start the CLI command.
  2. Log in to WeRead through the web reader.
  3. Open the target book from the bookshelf.
  4. Iterate chapters through the reader state.
  5. Convert chapter HTML into plain text.
  6. Write the output files.

Notes

  • The book title match is substring-based against the bookshelf text.
  • This tool assumes the requested book is already available in the logged-in WeRead account.
  • When --headless is used, the QR code is printed in the terminal only if the optional QR dependencies are installed.