End-to-end automation: from Zotero paper management to Obsidian deep-reading notes and wiki maintenance.
Zotero-Obsidian-Wiki-Bridge connects two powerful Codex skills into one seamless research workflow:
| Skill | Role |
|---|---|
| DeepPaperNote | Turns a single research paper into a high-quality Obsidian deep-reading note with evidence-first analysis, figure placeholders, and structured formatting |
| Obsidian Wiki Maintainer | Keeps your Obsidian research wiki consistent: auto-indexes, full-text search, linting, backfill, and topic/entity page generation |
Together they automate the most tedious parts of academic reading and knowledge management — so you can focus on thinking.
PDF / DOI / arXiv / Zotero
│
▼
DeepPaperNote
(evidence gathering, figure planning, note writing, lint + review)
│
▼
Obsidian Vault ──► Obsidian Wiki Maintainer
(index update, search refresh, conflict tracking)
Each skill can be installed independently via Codex skill installer:
npx skills add XuanjinZhu/zotero-obsidian-wiki-bridge/deeppapernote -a codex
npx skills add XuanjinZhu/zotero-obsidian-wiki-bridge/obsidian-wiki-maintainer -a codexOr clone the entire repository and use the scripts directly.
- Python >= 3.10
- PyMuPDF (
pip install PyMuPDF) — required for DeepPaperNote, optional for Wiki Maintainer - An Obsidian vault (recommended)
- Optional: Zotero for local paper management
Within Codex (or any compatible agent environment), invoke DeepPaperNote:
{给我这篇论文生成深度笔记
- DOI: 10.xxxx/xxxxx
After notes are written, run the maintainer:
python obsidian-wiki-maintainer/scripts/ingest_maintain.py \
--vault /path/to/your/vault \
--note /path/to/new/note.mdzotero-obsidian-wiki-bridge/
├── README.md
├── README.zh-CN.md
├── LICENSE
├── .gitignore
├── deeppapernote/ # Deep paper reading skill
│ ├── SKILL.md
│ ├── README.md / README.zh-CN.md
│ ├── CHANGELOG.md
│ ├── LICENSE
│ ├── pyproject.toml
│ ├── agents/openai.yaml
│ ├── assets/
│ ├── references/
│ ├── scripts/
│ └── tests/
└── obsidian-wiki-maintainer/ # Wiki maintenance skill
├── SKILL.md
├── README.md / README.zh-CN.md
├── CHANGELOG.md
├── LICENSE
├── pyproject.toml
├── agents/openai.yaml
├── references/
├── scripts/
└── tests/
- Evidence-first deep reading pipeline (not a summary generator)
- Model-led understanding: mechanism breakdown, key numbers, formulas, limitations
- Placeholder-first figure strategy — preserves figure context even when extraction is partial
- Obsidian-native output: paper-specific folder, Markdown note, local
images/directory - Zotero integration for local-library-first paper resolution
- Lint gate + final readability review before save
- Schema-driven: single
Research/AGENTS.mddefines your entire wiki layout - Auto-indexing:
index.md,log.md,conflicts.mdkept in sync - Full-text search with raw-source fallback layer
- Lint checks: broken links, orphans, missing frontmatter, stale entries
- Backfill: migrate legacy notes into canonical directory structure
- Query filing: persist valuable search insights into topic/entity pages
See each skill's own README for development setup, testing, and contribution guidelines.
- DeepPaperNote was originally created by dingdingcar. This fork preserves the original workflow and adapts it into the unified Zotero-Obsidian-Wiki-Bridge project.
- Thanks to the Obsidian and Zotero communities for building excellent knowledge management tools.
MIT — see LICENSE. Each sub-skill also carries its own MIT license.