Zhihu Article Fetcher & Obsidian Sync

Problem

Zhihu content is split across collections, personal activity feeds, and article pages. Copying manually loses structure, image references, and local searchability. zhihu-fetch-skill targets workflows that need a specific collection or like/bookmark history turned into local Markdown files with images and Obsidian-ready notes.

How it works

The skill separates tasks by scripts: a list script exports a collection as JSON, a batch script fetches article Markdown, downloads images, and maintains _progress.json for resumable runs. If the cookie expires, the API returns only summaries, or headless browsing is detected, it can fall back across API, DOM, stealth, and interactive browser strategies. Obsidian writing detects the Vault, places notes under existing Zhihu Collections categories or keyword-based folders, and preserves frontmatter such as title, author, url, and voteup. History workflows keep interaction metadata and deduplicate by URL to avoid duplicate imports.

Boundaries and caveats

This is not a generic web scraper; it is built around Zhihu login state, the z_c0 cookie, anti-bot checks, and Obsidian path conventions. Captchas, security verification pages, or repeated environmental failures may require manual login or network changes. Some images can fail due to Referer restrictions or expired links, and the article text may retain original URLs. Users should respect Zhihu terms, privacy rules, and copyright when processing content.