ScholarAIO v1.5.0

ScholarAIO v1.5.0 is a user-facing release focused on making the research library easier to browse, easier to verify, and easier to connect with real-world paper workflows.

Highlights

Local Library WebUI

The biggest visible change is the new read-only local WebUI:

scholaraio gui

It gives you a private browser interface for exploring your main library and proceedings records, with live refresh, filters, metadata and quality inspection, Markdown-rendered abstracts and conclusions, local math rendering, and inline PDF preview/fullscreen controls. It is designed for day-to-day library triage and reading without exposing library data to remote runtime scripts.

Source-grounded evidence search

ScholarAIO can now build and query line-addressable evidence chunks:

scholaraio index --chunks
scholaraio search --chunk "your query"

Results can include paper IDs, section titles, source snippets, and line ranges, so agent-assisted answers can point back to concrete paper text instead of relying on broad document matches.

Easier publisher PDF fetching

scholaraio fetch-pdf now supports DOI, publisher landing page URL, direct PDF URL, or title-based locators. It can use your current legal network access, supports direct campus-network mode, and can hand fetched PDFs into the normal ingest flow or refresh canonical PDFs for existing library records.

Paper2Any workflows from ScholarAIO

This release adds the optional Paper2Any integration path through scholaraio paper2any, including setup, status, smoke, MCP sidecar serving, and backend serving commands. The goal is to make paper-to-figure, slide, poster, and related conversion workflows easier to launch from the same research workspace.

Better setup diagnostics for optional tools

scholaraio setup check now gives clearer guidance for optional external services, including websearch, webextract, Paper2Any, Graphviz, Inkscape, MinerU, Docling, API keys, and configured endpoints. The diagnostics are meant to tell users what is ready, what is optional, and what command to run next.

Reliability and Privacy Improvements

• The local WebUI is hardened against malformed metadata, stale async responses, non-ASCII PDF filenames, large PDF buffering, stale toolbar/filter state, and remote runtime script loading.
• Local MinerU batch runs now keep image assets isolated per PDF, avoiding overwritten generic image names during batch conversion.
• Paper PDFs are preserved more safely during ingest and repair, including duplicate DOI recovery and forced replacement checks.
• Setup diagnostics now align reported Paper2Any and webtools endpoints with the actual runtime resolution order.
• Websearch and webextract HTTP health checks now respect the short setup-check timeout, so unavailable optional services do not slow down normal setup checks unnecessarily.

Important User-Facing Change

CLI help, status output, warnings, and service errors have moved toward English user-facing messages. If you have scripts that assert exact CLI text, check them after upgrading.

Upgrade Notes

For a normal package upgrade:

pip install -U scholaraio
scholaraio setup check

If you are upgrading from a pre-1.4 runtime layout, do not move data by hand. Use the explicit migration path first:

scholaraio migrate status
scholaraio migrate upgrade --migration-id upgrade-1.4.0 --confirm
scholaraio migrate verify --migration-id upgrade-1.4.0
scholaraio index --rebuild

Release Quality

This release was validated with the full test suite across the repository, package metadata alignment checks for v1.5.0, strict documentation build checks, wheel/sdist metadata checks, fresh wheel install smoke tests, and live optional-tool smoke paths where local dependencies were available.

See the full changelog for detailed changes: https://github.com/ZimoLiao/scholaraio/blob/main/CHANGELOG.md#150--2026-05-24