Methodology agent: bias-source calibration
0
Extend AI Review's methodology agent to score selection, performance, detection, and attrition bias on a 0-3 scale per STROBE/CONSORT axis, shown in the AI Review report beside its existing findings.
Roadmap
What we're building now, what's queued next, and what we plan for later. Last reviewed 2026-04-22.
Sign in to vote.
Now · 3
Figure/table extraction 2.0
0
Switch from text-only analysis to pdfplumber+pymupdf dual-pass. Extract figure captions alongside panels so the figures agent can score labelling consistency directly.
Paid-ads ramp-up (infra only)
0
Day 5 shipped the pixel loader + retargeting cohort builder. "Now" means: validate Meta Pixel + Google Ads fires in GTM Debugger under real consent. No spend yet.
Next · 3
Reviewer-rebuttal conversation loop
0
Let authors respond to a specific agent critique; re-run that single agent with the rebuttal in context. Keeps audit trail.
Per-paper citation graph
0
Show a force-directed graph of the works a manuscript cites and the papers around them, with OpenAlex backrefs.
Institutional dashboards
0
Tenant-scoped dashboards for university and lab tenants: AI Review and Pre-Check usage across their researchers, and the review axes that fail most often.
Later · 3
Human-in-the-loop reviewer network
0
Invite domain experts to co-review 10% of papers alongside our review engine. Training signal + credibility anchor.
Overleaf integration
0
Send a manuscript to AI Review directly from an Overleaf project, without exporting a PDF.
Mobile upload flow
0
Progressive upload + metadata capture on mobile. Currently blocked by PDF-upload UX on iOS Safari.