Dashboard Overview

  • Manuscript List (Homepage)- Displays all manuscripts processed through Alchemist Review, sorted by upload date. Users can search by Manuscript ID or title, and filter by upload date. Journal fit score is shown for each manuscript on the Manuscript List. Role based access control/Permissions determine which journals are visible to the user.
  • Manuscript Digest (Tab)- A consolidated summary view that organizes manuscript data into clear sections for faster analysis. Designed to minimize time spent navigating PDFs.
  • Journal Fit Score – Helps editors quickly assess how well a manuscript aligns with their journal’s scope and standards—before assigning reviewers or investing time in deeper review.
  • Flagged Items – To help with early triage, the number of critical severity (red), low severity (yellow) or neutral severity (grey) references are shown on the manuscript list.

Digest Sections

  • Journal Fit (detailed criteria)- The journal fit score is calculated based on the Ai’s evaluation across four core editorial criteria: Novelty, Relevance to Scope, Scientific Rigor, and Significance & Impact. 
  • Paper Type- Classifies the manuscript by category (e.g., research article, review paper).
  • General Summary- A plain-language overview of the manuscript at a “science magazine” level—intended for non-specialists.
  • Advanced Summary- A more technical synopsis tailored for domain experts, highlighting study design and findings in depth.
  • Methods- Outlines primary methods used in the paper, with inline access to AI explanations. Users can hover and select “Ask AI Chat” to learn more about specific techniques.
  • Key Concepts- AI-generated list of major terms and themes in the paper. These may differ from author-provided keywords. (Ongoing model tuning is improving accuracy.)
  • Author Contributions- Summarizes what the authors claim to contribute to the field, drawn directly from the manuscript text.
  • Key Issues & Proposed Remedies- Identifies weaknesses in the research (e.g., methodological gaps) and AI-suggested remedies. Note: may over-flag “lacking state-of-the-art methods”; refinement in progress.
  • Writing Quality & Readability- Analyzes text clarity and flow, flagging issues like run-on sentences or structural confusion.

Citation Evaluation

  • Citation Analysis (Digest Summary)- Summarizes checks from the Citation Evaluation tab—flags missing DOIs, invalid sources, or irrelevant references.
  • Citation Evaluation Tab- Detailed reference-level assessment including:
  • DOI Validity – Confirms correct DOI formatting and matching.
    • Source Validity – Verifies citations link to legitimate publications.
    • Relevance Score – Rates how appropriately each reference supports the paper
    • Retraction Flags – Highlights retracted papers or authors with prior retractions.
    • Severity Filters – Classifies issues as Critical, Minor, or Pass.
    • Interactive Table – Clickable list linking to full references and in-text citations.
  • Author Representation- Analyzes the proportion of self-citations and author repetition—potential signal for citation rings.
  • Journal Representation- Shows how often each journal is cited, helping assess whether references align with journal scope and audience.
  • Recency Analysis- Charts publication dates of cited references, allowing filtering by timeframe (e.g., last 5, 10, or 25 years).

AI Chat

AI Chat Panel – An interactive assistant built into the dashboard for manuscript-level inquiry. It is built on Open Scholar and is not limited to information in the manuscript itself. The chat history is saved per manuscript.

Users can:

  • Ask about methods, data reliability, or conceptual meaning.
  • Access pre-populated prompts via hover icons beside digest items.
  • Retrieve context from manuscript text, digest data, and external sources via Open Scholar.
  • View per-manuscript conversation history for auditability.

Inline Feedback 


 Allows users to rate AI-generated content (Thumbs up/down) and provide qualitative comments (e.g., “too much information,” “irrelevant”). Feedback is reviewed by Hum’s product team regularly.

 

Administrative Functions

Permissions Management – Controls journal access for users; editors may have restricted visibility while staff can access multiple titles.