See also our Privacy policy.

Policy-first publishing

AcademiaX Press treats publishing as an engineering discipline: a paper is not “done” because it looks good; it is “done” when it satisfies a set of explicit, testable requirements. Policies apply across journals, books, and the agent pipeline so that quality is consistent as volume scales.

  • Traceability: a reader should be able to trace claims back to sources, versions, and decisions.
  • Repeatability: given the same sources and constraints, the pipeline should be able to reproduce outputs deterministically.
  • Bounded iteration: revision loops have maximum cycle thresholds to prevent endless churn.

Citations-first

Drafts must cite only provided sources. Where a claim cannot be supported by the internal evidence base, it is either removed or (optionally) escalated to grounding.

What this means in practice

  • We prefer “show the source” over rhetorical certainty.
  • When sources disagree, the paper must say so and cite both sides.
  • When a fact is unknown, the paper must label uncertainty rather than speculate.

Verification

  • DOIs are verified via Crossref; mismatches are discarded.
  • Tables store traceable references to source DOIs.
  • Outputs are versioned and stored as URIs in object storage.

Grounding tiers

Some environments enable additional grounding tiers (retrieval and evidence tracing) for high-stakes claims. When enabled, the pipeline records evidence pointers and surfaces them for review.

Peer review and decision policy

Reviews are structured, repeatable, and actionable. Critic agents produce directives (what to change and why); an Editor‑in‑Chief agent aggregates scores and enforces a maximum number of revision cycles.

  • Accept when coverage, methodology, and evidence discipline meet the bar.
  • Revise when issues are fixable within bounded cycles.
  • Reject when the scope is mismatched, evidence is insufficient, or the paper cannot be fixed within limits.

Authorship, attribution, and disclosures

Authors are responsible for the final content. If AI assistance is used (for drafting, editing, or review), the work should disclose use at an appropriate level of detail. We also encourage ORCID integration where available.

  • Authorship: list contributors who made substantive scholarly contributions.
  • Acknowledgements: credit tools, datasets, and non-author contributors appropriately.
  • Conflicts of interest: disclose financial, personal, or competitive conflicts.

Corrections, updates, and retractions

We version outputs. When mistakes are found, we prefer transparent corrections: publish an updated version, document what changed, and preserve access to prior versions when appropriate.

  • Correction: for discrete factual or citation errors.
  • Update: for substantial improvements, expanded evidence, or new results.
  • Retraction: for misconduct, irreparable errors, or violations of evidence policy.

Data, artifacts, and reproducibility

When a paper relies on data, code, or tools, authors should provide access where possible. If access cannot be provided (privacy, licensing), papers should document the limitation clearly.

  • Provide datasets or dataset descriptors and licenses where applicable.
  • Provide reproducible methods and parameter choices (including model settings where relevant).
  • Ensure figures/tables can be traced to sources or generated artifacts.

Security, privacy, and abuse prevention

We protect accounts, limit access by role (RBAC), and apply abuse resistance for user-submitted forms. If you discover a security issue, contact security@academiaxpress.com.

Cookies

We use essential cookies to keep the site functional and to remember preferences. Analytics can be enabled in enterprise deployments.

Where to go next