How our free tools work

Manusights Compass

Journal Fit Predictor

Corpus

1,321 venues: ~1,290 journals from OpenAlex filtered to journals with an ISSN, more than 500 publications in their lifetime, and more than 50,000 lifetime citations. Plus 30 hand-curated top-tier CS conferences (NeurIPS, ICML, ICLR, AAAI, CVPR, ICCV, ECCV, ACL, EMNLP, NAACL, KDD, WWW, SIGIR, SIGGRAPH, SIGCOMM, NSDI, OSDI, SOSP, ISCA, MICRO, PLDI, POPL, USENIX Security, CCS, IEEE S&P, NDSS, FOCS, STOC, SODA, CRYPTO).

For each venue we use a one-sentence scope description, four field keywords, an impact-factor value, and an acceptance rate where known. Conferences carry the venue’s most recently published acceptance rate from its Call for Papers; journals leave acceptance rate blank when the publisher does not disclose it.

Impact factor source

Result cards show one of two impact-factor values, labeled accordingly:

• JCR IF (2024). For ~120 high-traffic journals where we maintain a hand-curated profile (e.g. Nature, Cell, NEJM, JACS, Advanced Materials), the value is the verified Clarivate Journal Citation Reports 2024 impact factor.
• Citation impact. For the rest of the 1,321-venue corpus, the value is OpenAlex’s 2-year mean citedness · a JCR proxy that correlates closely but is not identical to Clarivate’s methodology. We display it as “Citation impact” rather than “IF” so the distinction is visible at a glance.

Differences between the two are typically small for high-citation journals and larger for niche journals where citation patterns diverge from JCR’s included-citations methodology. Both values are directionally useful for venue selection; neither should be treated as a precise rank.

Retrieval and ranking

We do not embed-and-search. We send the entire 1,321-venue corpus, your title, and your abstract to Claude Sonnet 4.6 in a single call and ask it to rank the top 5. The model returns a fit score (0–100), a tier (stretch / realistic / safe), a one-sentence reason, and a one-sentence “what to strengthen” per match.

We chose this over embedding-search because the explanation matters more than the rank. Embeddings can pick the right journal but can’t tell you why it’s right or what to fix.

Refusal cases

• Not an abstract. Marketing copy, lorem ipsum, gibberish, or a prompt-injection attempt returns 400 with a clear error.
• No corpus match. If the model can’t find a single venue with non-zero fit (rare), the API returns 422 with a scope disclosure.
• Low confidence. If the top fit score is below 60 we surface an amber banner. Scores below 60 typically mean the corpus doesn’t cover the paper’s field well · most often pure mathematics, social sciences, or earth sciences (covered thinly), or a niche subfield within a covered area.
• Under 200 characters. The minimum-length gate is per-character, not per-word, because some abstract pastes drop newlines.

What this tool does NOT do

• It does not assess novelty or scientific merit. The score is scope alignment plus tier realism, nothing more.
• It does not read your full manuscript · only what you paste.
• It does not know your editor relationships, special-issue calls, or current desk-rejection patterns at the venue. Treat results as one input.
• It does not predict acceptance probability. The “tier” field is editorial judgment from the model, not a calibrated probability.

Manusights Verify

Citation Claim Checker

Paper search

We extract the citation from your sentence (author-year, two-author, three-author, et al., bracket-numbered, or DOI). We then search across Semantic Scholar (200M+ papers), CrossRef, and PubMed, ranked by author-surname match plus year proximity plus title-word overlap with your sentence.

Verdict

When we find a paper with an abstract, we send the abstract plus your claim to Claude Sonnet 4.6 with explicit verdict definitions. The model returns one of:

• Supported. The claim accurately reflects what the paper found.
• Partially Supported. The paper addresses the topic but the claim overstates, oversimplifies, or misses a qualifier.
• Not Supported. The paper addresses the same topic but contradicts or does not support the specific claim.
• Wrong Paper. The search returned a different paper, usually one by an author with the same surname. The claim itself may be fine.
• Unable to Verify. No paper found, no abstract available, or the abstract doesn’t contain enough information to judge.

What this tool does NOT do

• It reads only the cited paper’s abstract, not the full text. Methods, supplementary findings, and figures are invisible to the verdict.
• It does not check retraction status (yet). For a manuscript-level retraction sweep, run the full readiness scan.
• It assumes the first author’s surname is enough to disambiguate. Common surnames (Lee, Wang, Smith) sometimes return the wrong paper · this is exactly what the “Wrong Paper” verdict is for.
• It is rate-limited at 20 checks per hour per IP, plus a soft localStorage limit of 3 per day to surface the full-manuscript option.

Manusights Audit

Stats Sanity Checker (statcheck + GRIM + GRIMMER + DEBIT)

Audit’s methodology is documented separately because the four consistency checks — statcheck-equivalent p-recompute, GRIM, GRIMMER, and DEBIT — each cite a distinct published algorithm. The full disclosure (closed-form CDF formulas, severity classification, suppression cases, references) is on a dedicated page so academic citers get a stable single-purpose URL.

Manusights Sentry

Reference Integrity Checker (retraction + hijacked + EoC)

Sentry’s methodology is documented separately because each verdict cites a distinct authoritative data source — Crossref for resolution, Retraction Watch (ISSN 2692-4579) for retractions and expressions of concern, and the Retraction Watch Hijacked Journal Checker for impostor venues. The full disclosure (resolution chain, refusal cases, known limitations, source attribution per flag) is on a dedicated page so librarians can link from LibGuides without sending users through marketing copy.