lark-doc-exporter 0.3.0

Export Feishu/Lark docs with synced blocks to Markdown and themeable PDF

lark-doc-exporter

Export Feishu/Lark docs with synced blocks expanded into:

• Markdown with localized images
• Themeable locally rendered PDF by default
• Optional native Feishu PDF mode with AI-footer post-processing

Requirements

• lark-cli configured with a user session
• Python 3.14
• For --pdf-mode rendered: a working Chrome/Chromium runtime, or the ability to install one with uvx --from playwright playwright install chromium
• For --pdf-mode native: PyMuPDF available in the runtime environment (bundled with this package dependency as pymupdf)

If you use uvx / uv tool install, uv can provision the required Python for the tool environment automatically.

Quick Start

uv tool install lark-doc-exporter
lark-doc-exporter doctor

If you want rendered PDF output and Chromium is missing, prepare it once:

uvx --from playwright playwright install chromium

Then run an export:

lark-doc-exporter \
  --doc "https://dynamia-ai.feishu.cn/wiki/WEgBwqGYOiBoQikRzjncvJDonAg" \
  --output-dir exports/demo \
  --formats markdown,pdf \
  --theme default

One-off Run

If you do not want a persistent tool install:

uvx lark-doc-exporter \
  --doc "https://dynamia-ai.feishu.cn/wiki/WEgBwqGYOiBoQikRzjncvJDonAg" \
  --output-dir exports/demo \
  --formats markdown,pdf \
  --theme default

Install As A Tool

After installing the released package, companion-skill operations stay the same:

lark-doc-exporter skill install --dry-run
lark-doc-exporter skill install

Auto mode installs the companion skill into every detected supported host:

• Codex: ~/.agents/skills/lark-doc-exporter
• Claude Code: ~/.claude/skills/lark-doc-exporter

Use --host codex, --host claude, or --host all to target specific hosts. --dry-run previews the install plan and target directories without writing files. Use --force only when you intentionally want to replace an existing unmanaged target directory.

Chromium Setup

• doctor reports whether lark-cli is ready and whether Chromium is available for rendered PDF mode.
• If you already have a browser binary, you can point the exporter at it with LARK_DOC_EXPORTER_CHROMIUM=/path/to/chromium.
• If LARK_DOC_EXPORTER_CHROMIUM is set to a missing path, the command fails explicitly instead of silently falling back.

Output

• markdown keeps the localized Markdown file in the output directory.
• pdf uses either local HTML/CSS + Chromium (--pdf-mode rendered) or Feishu native PDF plus footer handling (--pdf-mode native).
• images/ contains same-run localized image assets used by the Markdown/PDF.

Native PDF Mode

If you want Feishu native PDF layout instead of the local HTML/Chromium path:

lark-doc-exporter \
  --doc "https://dynamia-ai.feishu.cn/wiki/BVXXwgzbZivjQZkr7jmcsGcinGh" \
  --output-dir exports/native \
  --formats pdf \
  --pdf-mode native

Native mode rules:

• only the PDF branch changes; markdown stays on the current markdown pipeline
• explicit non-default --theme / --css are rejected
• success states are removed and not_found
• failure states emit warnings and keep <stem>.native-raw.pdf for inspection

Themes

Built-in themes:

• default
• company

You can also layer custom CSS on top:

lark-doc-exporter \
  --doc "https://dynamia-ai.feishu.cn/wiki/WEgBwqGYOiBoQikRzjncvJDonAg" \
  --output-dir exports/company \
  --formats pdf \
  --theme company \
  --css /path/to/your-company-print.css

Development / Unreleased

Use the Git URL or a local checkout only when you intentionally need unreleased code:

uvx --from git+https://github.com/spencercjh/lark-doc-exporter lark-doc-exporter doctor

git clone https://github.com/spencercjh/lark-doc-exporter
cd lark-doc-exporter
uv sync --python 3.14 --group dev
make fmt
make ci

# Optional runtime/environment check (not part of required CI)
uv run lark-doc-exporter doctor

Notes

• The current implementation still uses a temporary Feishu doc to translate the expanded XML into fresh Markdown.
• The temporary doc is deleted by default after the Markdown export step. Use --keep-temp-doc only when you need to inspect that intermediate document.
• Feishu image authcode URLs expire quickly, so image localization happens in the same run as the Markdown export.