Changelog ChatAcademia

Stay up to date with all the latest updates, improvements, and new features we ship to help you do better research. We’re constantly improving ChatAcademia for academics and researchers.

Home

© 2026 ChatAcademia

Database-Grounded Agent Checks

Database-Grounded Agent Checks

More beta agents now start from academic database context, while Claim Evidence and Hallucination Checker make paper links and DOI checks clearer before asking you to review anything manually.

Improvements

  • Use database selectors in Analysis Foundry, Seminar Architect, and Study Steward so those agents can ground plans in selected academic sources
  • Ask agents to search related papers before carrying out academic tasks, with DOI and reference resolution prioritized when available
  • Open Claim Evidence papers directly from evidence cards, strongest-evidence snippets, overview lists, and tool previews
  • Audit Hallucination Checker references with DOI.org metadata alongside academic database and provider lookups
  • Keep likely hallucinated references out of the real-paper link count so unrelated matches are not shown as verified papers

Analysis Foundry Beta Agent

Analysis Foundry Beta Agent

Analysis Foundry helps researchers turn datasets, variables, and hypotheses into practical analysis plans with model choices, reproducibility checks, and reporting guardrails.

Improvements

  • Start the new Analysis Foundry beta agent from the app sidebar and agent discovery surfaces
  • Profile variables, measurement types, missing information, and data quality risks before analysis
  • Map research questions to analysis strategies, estimands, assumptions, alternatives, and decision rules
  • Draft statistical model plans with covariates, diagnostics, missing-data handling, robustness checks, and interpretation limits
  • Build reproducibility checklists covering data provenance, code structure, environments, versioning, and handoff evidence
  • Review data readiness, strategy, model, reproducibility, and reporting modules in a dedicated project panel beside the chat
  • Open the new website landing page for Analysis Foundry from agent discovery, header, footer, and homepage surfaces

More Stable Sign-In Callbacks

More Stable Sign-In Callbacks

Sign-in and sign-up redirects now hydrate more consistently on the loading screen, reducing rare callback failures during account access.

Improvements

  • Keep the app loading screen stable between the server-rendered page and the browser
  • Reduce rare sign-in callback crashes on mobile browsers
  • Preserve the existing ChatAcademia loading experience while improving startup reliability

Seminar Architect Beta Agent

Seminar Architect Beta Agent

Seminar Architect helps instructors turn a course, seminar, or workshop idea into a practical teaching plan with modules, assessments, rubrics, timelines, and reusable feedback comments.

Improvements

  • Start the new Seminar Architect beta agent from the app sidebar and agent discovery surfaces
  • Map learning outcomes into modules, activities, assessment touchpoints, inclusion notes, and alignment metrics
  • Build assessment plans with weighting, feedback windows, integrity checks, accessibility considerations, and readiness signals
  • Draft analytic rubrics and reusable feedback banks that instructors can review before using with students
  • Review timeline, workload, rubric, assessment, and feedback outputs in a dedicated project panel beside the chat
  • Open the new website landing page for Seminar Architect from agent discovery, header, footer, and homepage surfaces

More Reliable Reference Manager Saves

More Reliable Reference Manager Saves

Saving papers from chat to Zotero or Mendeley is more forgiving when a paper has useful metadata but no parseable full text. Agents now keep the reference details instead of failing the whole save.

Improvements

  • Save papers to Zotero or Mendeley with full metadata even when no public full text can be parsed
  • Attach the resolved PDF and full text only when the document can be fetched and parsed successfully
  • See clearer guidance when PDF Vector cannot fetch or process a document URL