How SubmitPilot works, and what it cannot tell you.

1. Finding candidate journals

Your topic, title or abstract is reduced to its most distinctive terms. If you upload a manuscript, its keywords are used. We search OpenAlex for journal articles and reviews published in the last four years that match those terms, take up to 200 of the most relevant, and group them by the journal that published them. Preprint servers, conference proceedings and repositories are excluded.

2. Scoring fit

Each journal receives a relative fit score from three parts:

• Evidence (50%): how many of the matching papers it published, with higher-ranked matches counting more. Very large journals are damped slightly so that volume alone does not win.
• Title overlap (25%): how closely the titles of those papers match your terms.
• Topic profile (25%): overlap between your terms and the journal’s main topics in OpenAlex.

The score orders results; it is not a probability. It is shown as scope fit: strong (score 60 or more and at least three matching papers), partial, or weak.

3. Desk-rejection risk

Risk is a simple points system. A weak scope fit adds two points and a partial fit one. A highly cited journal adds one. With a manuscript, missing core sections add one, and a short reference list submitted to a top-half journal adds one. Three or more points is higher risk, two is moderate, fewer is lower. The reasons are listed with every rating, so you can judge them yourself.

4. The six lists

• Best match: up to five ranked journals with at least partial fit and no cautions.
• Q1 to Q4: when SCImago Journal Rank data is loaded, these are SJR best quartiles. Otherwise they are impact tiers: the matched journals are sorted by two-year mean citedness in OpenAlex and split into four equal groups. Impact tiers are relative to your results, are not official quartiles, and are labelled “Tier” rather than “SJR”.
• Unranked: journals without ranking data. Some are new or regional; some are questionable.

Citation metrics shown

• SJR score and SJR quartile from SCImago Journal Rank (Scopus data), refreshed yearly.
• Cites per doc (2 years) from SCImago: citations received in a year by papers the journal published in the previous two years, divided by the number of those papers. It is calculated the same way as an impact factor, but from Scopus rather than Web of Science, so the numbers differ.
• Mean citedness from OpenAlex, a similar two-year measure from open data.

We do not show the Journal Impact Factor or CiteScore. Both are licensed by their owners (Clarivate and Elsevier) and cannot be republished freely. Check them on Journal Citation Reports or Scopus Sources.

5. Cautions

A journal is flagged for a closer look when:

• it is in neither the OpenAlex core-source list nor DOAJ (and not in SJR when loaded);
• its listed publication fee is $1,000 or more while its papers are rarely cited;
• its yearly output grew at least threefold in three years to over 1,500 papers;
• no papers have been indexed for more than a year.

A caution is a reason to check, not proof of a problem. Our guide to predatory journals lists the checks to make.

6. Submission requirements

For the journal you choose, we start from the publisher’s standard requirements (for example MDPI’s back matter statements or IEEE’s Index Terms). We then look for the journal’s author-guidelines page. If it can be read and an AI model is available, the model extracts the reference style, limits and required statements from that page, and each item is labelled with its source. Many publishers block automated access; in that case the publisher defaults are shown and labelled as such.

7. Manuscript checks

Uploaded files are converted to plain text and scanned for section headings, abstract and keyword lines, in-text citation patterns and the reference list. The checks are rules, not judgements of scientific quality. PDFs can lose their heading structure, so a .docx gives more reliable results.

8. AI models

The “what is missing” analysis and guideline extraction use open-weight models (such as GLM, Qwen or DeepSeek) through OpenAI-compatible APIs. The model sees the journal’s recent paper titles and topics plus your topic or abstract. It is instructed not to invent facts about journals. When no model is available, a rule-based comparison of terms is shown instead and labelled.

Limits

• OpenAlex coverage is broad but not complete, and journal metadata can lag behind changes.
• Scope fit is based on titles and topics, not full texts.
• Nothing here predicts acceptance. Editors weigh novelty, quality and timing that no tool can see.