05 Policies
Rules of the house.
How we decide what to publish, how every kind of object is reviewed, and what happens when something needs to change. Our policies follow the Core Practices of the Committee on Publication Ethics (COPE).
Editorial independence
Decisions to publish are made by editors, on the merit of the object and the outcome of its review. Nothing else decides them.
- The people who prepare quotes and invoices take no part in editorial decisions. Paying for a quote or a service never buys acceptance.
- Every quote says in advance what is charged if an object does not pass review.
- Editors and reviewers declare any conflict of interest and step aside when they have one.
- The owners of Metamahal do not intervene in decisions about individual objects.
Peer review and quality review
Every object we publish is reviewed before publication. The kind of review depends on the object: a book is read by experts in its field; a dataset is opened and checked against its documentation; a model is checked against its card, the rights to its training data and an evaluation that can be repeated.
| Room | Object | Review |
|---|---|---|
| 01 | Books and chapters | Peer review of the full manuscript by at least two independent experts |
| 02 | Conference proceedings, posters and slides | Check of the conference and its review; independent review where that is missing |
| 03 | Reports, theses and working papers | Independent expert review for reports; confirmation of the award for theses |
| 04 | Preprints and peer reviews | Screening for preprints; consent of reviewer and author for published reviews |
| 05 | Datasets and data papers | Data quality review; data papers peer reviewed like articles |
| 06 | AI models and model cards | Review of the model card, training-data rights and a repeatable evaluation |
| 07 | Software, notebooks and protocols | The code runs, tests pass, the release is archived and licensed |
| 08 | Learning and audiovisual works | Subject-expert review, accessibility, consent and permissions |
| 09 | New kinds of objects | Agreed object by object, and published alongside it |
How reviewers are chosen
- For their expertise in the subject and in the kind of object — a data reviewer for data, a software reviewer for code.
- Without conflicts: no recent collaborators of the authors, no one from the same institution, no competing interests.
- We may invite reviewers whom authors suggest, but never only them, and we confirm every reviewer’s identity independently.
Review models
Single-anonymous, double-anonymous or open review, agreed for each object or series. Open review — where names and reports are published — happens only with the consent of everyone involved.
Decisions
Accept, revise or decline, always with reasons. Authors see every review report. Preprints and working papers are screened rather than peer reviewed, and are labelled that way.
Publication ethics
We follow COPE’s guidance across its ten Core Practices:
- Allegations of misconduct are investigated following COPE’s guidance, involving institutions where needed.
- Authorship and contributorship — see below.
- Complaints and appeals — see below.
- Conflicts of interest — see below.
- Data and reproducibility: texts say where their data and code are; for datasets, software and models the object itself is the data or code, and is reviewed as such.
- Ethical oversight: work involving people or animals states its approval and consent; people who appear in audiovisual works agree to publication; knowledge held by a community needs the community’s informed consent.
- Intellectual property: licences declared, third-party material cleared, similarity checks on every text.
- Publisher management: ownership, management and these policies are public on this site.
- Peer review processes — see above.
- Post-publication discussions and corrections — see below.
Authorship and contributorship
- Everyone listed as an author or creator made a substantial contribution and approves the published version.
- Contributions are recorded with the fourteen roles of the CRediT taxonomy, so that curation, software and analysis are credited as clearly as writing.
- AI tools cannot be authors or creators: as COPE puts it, they “cannot take responsibility for the submitted work”.
- Changes to the list of authors after submission need the written agreement of everyone on it.
Conflicts of interest
Authors, reviewers and editors declare financial and non-financial interests that a reader might think relevant. Authors’ declarations are published with the object. Anyone with a conflict about an object takes no part in decisions about it.
Use of AI tools
Authors and creators
- You may use AI tools. Say which tool and version you used, for what and where — in the methods, the acknowledgements or the record’s description — as COPE asks.
- You remain responsible for everything in the work, including anything a tool produced.
- Images, audio and video generated by AI are labelled as such in the work and in its record.
Reviewers and editors
- Submissions are confidential. Reviewers and editors do not upload them to AI tools that keep or learn from what they are given.
Metamahal
- We use software for some checks — similarity, file validation, formatting — and say so in the review report. Decisions are made by people.
- AI models we publish must come with a model card; see AI models and model cards.
Corrections, versions and retractions
The published record is permanent, and we change it only in the open.
- Correction — an error that does not change the conclusions is corrected, with a dated notice linked to the original.
- New version — datasets, software and models are updated by publishing a new release with its own identifier. The new one points back to the old (
IsNewVersionOf) and the old one stays available. - Expression of concern — published while a serious question is investigated.
- Retraction — following COPE’s retraction guidelines. The object stays visible, clearly marked as retracted, with the reason.
- Removal — only for legal reasons or serious risk. The identifier then leads to a page that says what was removed and why.
Every notice is linked to the original in the metadata, so services that read it — including Crossref’s Crossmark — can show readers the current status.
Preservation and access
An identifier is only as permanent as the thing it points to.
- Before our first publication we will arrange long-term preservation with a recognised digital preservation archive, so that published objects remain available even if Metamahal cannot continue. This page will name the archive once the agreement is signed.
- We keep checksums and more than one copy of every published file.
- Landing pages stay online. If an object is withdrawn, its identifier leads to a notice — never to a dead link.
- Every object is published open access, free to read without registration.
Licences and copyright
Authors and creators keep their copyright and grant us the right to publish. Objects are published under open licences — by default CC BY 4.0 for texts and media, CC0 or CC BY 4.0 for data, an OSI-approved licence for software, and an open licence named in the model card for AI models. The licence is stated on the landing page, in the object and as a URL in its record. Our metadata records are released under CC0.
Ownership and management
The Principles of Transparency and Best Practice in Scholarly Publishing (COPE, DOAJ, OASPA and WAME) ask that ownership and management be clearly shown. Here they are.
- Owner
- Meta Mahal, a partnership firm registered in Telangana, India
- Registered address
- 4-178/30/2, Eshwaripuri Colony, Road No. 5, Nagole, Hyderabad, Telangana 500068, India
- Management
- Aashrith Katakam
Complaints and appeals
- Write to us through the contact form or at info@metamahal.org. We acknowledge every complaint and tell you who is handling it and when to expect an answer.
- To appeal an editorial decision, explain why you think it was wrong. An editor who took no part in the original decision reviews it.
- Concerns about a published object — errors, misconduct, rights — are handled under the corrections and ethics policies above.
Version 1 · 23 October 2026