Installation¶
Requirements¶
- Python 3.10+
- Git
Install from PyPI¶
# Core installation
pip install scrinium
# Full installation (import + pdf + office + draw)
pip install "scrinium[full]"
Then run:
scrinium setup
Install from Source¶
git clone https://github.com/wszqkzqk/scrinium.git
cd scrinium
# Core only (search, export, audit)
pip install -e .
# Full installation (import + pdf + office + draw)
pip install -e ".[full]"
Use the source install path when you want to inspect the codebase, edit the package locally, or contribute changes upstream.
Upgrading from ScholarAIO (pre-fork)¶
Scrinium is a hard fork of ScholarAIO; the git history is continuous, so an existing checkout upgrades in place. Data (data/, workspace/, config.local.yaml) is untouched.
cd /path/to/scholaraio # your existing checkout
git remote set-url origin git@github.com:wszqkzqk/scrinium.git
git fetch origin && git switch main && git pull --ff-only
pip uninstall -y scholaraio # drop the old editable install
rm -rf scholaraio.egg-info
pip install -e ".[full]" # install scrinium (new CLI name)
scrinium --version # 3.0.0
scrinium index # one-time index migration (schema v2: drops legacy vector tables)
Compatibility fallbacks (with a deprecation warning):
SCHOLARAIO_CONFIGis honored whenSCRINIUM_CONFIGis unset~/.scholaraio/config.yamlis used when~/.scrinium/config.yamldoes not exist
See Migrating from ScholarAIO for the 3.0 breaking changes and cleanup steps (removed LLM/embedding features, manual deletion of old model artifacts).
Optional tidy-up: rename the checkout directory to scrinium (recreate the venv afterwards, or keep a scholaraio -> scrinium symlink so old venv shebangs keep resolving); update any ~/.agents/skills symlinks that point at the old path; mirror the new code to other remotes you pull from (e.g. git push <mirror> main).
Optional Dependencies¶
| Extra | What it adds |
|---|---|
pdf |
PyMuPDF-based PDF fallback and long-PDF utilities |
import |
Endnote / Zotero import |
office |
DOCX / PPTX / XLSX ingest and inspection |
draw |
Mermaid and Inkscape-powered diagram generation |
full |
Core research workflow extras: import + pdf + office + draw |
dev |
Development tools (pytest, ruff, mypy) |
The embed and topics extras were removed in 3.0 together with all in-framework embedding/LLM features; no model download is ever required.
Setup Wizard¶
Run the interactive setup wizard to configure API keys and directories:
scrinium setup
Or check what's already configured:
scrinium setup check
setup check is the most complete initial diagnostic surface. It covers:
- core setup items: dependency groups,
config.yaml, MinerU / Docling availability, parser recommendation,contact_email, and directory state - optional advanced items: Semantic Scholar API key and Zotero API key
Current setup guidance prefers MinerU first whenever a MinerU path is available (local service or mineru-open-api + token). Docling and then PyMuPDF remain the fallback chain when MinerU is not usable or when the user explicitly prefers a lighter parser path.
Cost transparency:
MINERU_TOKEN: free to applycontact_email: freeSemantic Scholar API key: optional; most endpoints work anonymously, but some require a keyZotero API key: optional; Scrinium's current Web API import path expects it, while localzotero.sqliteimport does not
Agent Setup¶
If you want to know which path to use for Claude Code, Codex, OpenClaw, Cursor, or other agents, see:
That guide separates:
- opening this repository directly
- registering Scrinium for use from another project
- choosing between native skills and plugins