Articles and RSS/Atom

The article reader prepares deterministic text evidence from local documents, feeds, and explicitly approved public article fetches.

Status: shipped

Entry points: voidscape.py, article.py

Inputs#

• Local Markdown, plain-text, or HTML documents.
• Local RSS or Atom files, up to 100 entries.
• Non-video http/https article URLs after preview and explicit fetch approval.

Video-host URLs remain on the video reader. Complex publisher templates, subscriber-only pages, and browser-authenticated article extraction are not promised.

Approval boundary#

Local documents and feeds require no network. Fetching an article URL transfers the URL/request to the remote server and is blocked until the previewed job is approved (--allow-cloud in the guided CLI or --allow-fetch in article.py). The reader does not reuse browser login state.

Guided use#

voidscape inspect "feed.xml"
voidscape preview "feed.xml"
voidscape read "feed.xml" --workdir article-evidence

Raw commands follow manifest -> probe -> estimate -> run; run writes manifest.json and ordered text files under entries/.

Citation contract#

• A single article uses [article N], normally [article 1].
• RSS/Atom items use [entry N] in manifest order.

Do not invent page numbers, timestamps, publication metadata, or text the extraction did not produce. Canonical details: article and RSS reader.