Literature monitoring
Research alerts built around the question you are actually following
Create a scheduled literature watch from your research profile, selected library papers, a plain-language brief, or a combination of all three.
A useful research alert needs more than a recurring keyword. Terminology changes, adjacent methods matter, and the newest item is not always the most useful item. miscite lets you define the evidence behind an alert, choose how broadly it should search, preview the result, and set a cadence that matches the pace of the topic.
Each digest is organized into Latest Research, Most Relevant, and Classics. That separation makes different reading decisions explicit: scan what just appeared, revisit strong matches to the alert context, or recover established work that may still frame the question. Past recommendations are tracked so the alert does not simply repeat the same feed.
Capabilities
What makes an alert specific
The alert builder separates source context, search behavior, schedule, and preview so each decision can be checked.
Context from the work you already do
Profile-based alerts can draw on your Interest Tags and connected public-source context. Library-based alerts can anchor discovery around particular saved papers. A custom brief lets you state a method, phenomenon, population, or research question in ordinary language.
Search behavior you can choose
Semantic retrieval emphasizes related meaning, lexical retrieval emphasizes shared wording, and hybrid retrieval balances the two. Optional Web of Science filters can limit results by selected indexes and JIF or JCI quartiles when that boundary is appropriate for the alert.
A digest structured for different reading jobs
Latest Research highlights recent matches, Most Relevant prioritizes fit to the alert, and Classics preserves influential or established context. Recommendations are stored with their section and search context so the digest remains more informative than an undifferentiated list.
Refinement informed by engagement
Searches, library saves, alert settings, and recommendation feedback can contribute signals to research updates. When enough dismissal feedback exists, a refinement step can use those patterns to keep, deprioritize, or remove candidates. If refinement fails, the underlying search results remain available.
Process
How to set up a research alert
The four-step builder makes it possible to inspect the alert before committing to recurring delivery.
-
Choose the alert sources
Select profile context, one or more library items, and/or a custom brief. Combining them can make a broad interest more concrete without forcing it into a single keyword string.
-
Tune retrieval and quality boundaries
Pick hybrid, semantic, or lexical retrieval. Enable WoS filtering only when the research question calls for it, then select any relevant index or quartile limits.
-
Set a realistic cadence
Choose daily for a narrow fast-moving watch, weekly for a regular review cycle, or monthly for a slower or broader topic. Weekly and monthly schedules include a chosen delivery day.
-
Inspect the live preview
Review sample results grouped as Latest Research, Most Relevant, and Classics. Adjust the source brief or filters if the preview is too broad, too narrow, or centered on the wrong concept.
Best fit
When scheduled alerts are a good fit
Alerts work best for questions that deserve repeated attention but not a fresh manual search every day.
Fast-moving specialties
Follow a narrow method, dataset, intervention, or policy area where recent work can change an active project.
Long-running reviews
Monitor an evolving literature between formal search updates while keeping the original review protocol separate.
Ideas anchored to key papers
Use one or more library references as the conceptual starting point when terminology alone is not a reliable description.
Boundaries
What research alerts cannot guarantee
An alert is a discovery aid. Its ranking and coverage should be evaluated against the standards of the project.
- No recurring search can guarantee that every relevant paper is indexed, described with complete metadata, or retrieved for a particular query.
- Semantic similarity is not the same as methodological quality or applicability. Preview the feed and read the source before acting on a recommendation.
- Engagement-aware refinement activates only when enough feedback is available and is designed to fall back to the unrefined candidate set when the refinement step cannot complete.
- Email delivery depends on the deployment's mail configuration. Digests and their recommendations remain part of the signed-in research workspace.
FAQ
Questions about research alerts
Can I create an alert without connecting a public profile?
Yes. An alert can use selected library references, a custom plain-language brief, or both. Profile context is optional, so a narrowly defined project does not need to inherit every topic from a broader researcher profile.
What is the difference between semantic, lexical, and hybrid alerts?
Semantic retrieval looks for related meaning even when wording differs. Lexical retrieval prioritizes direct term overlap. Hybrid retrieval combines both signals. The preview helps you decide which behavior fits the vocabulary and breadth of the topic.
How often can a research alert run?
The alert builder supports daily, weekly, and monthly schedules. A narrow, fast-moving topic may justify daily review, while weekly or monthly delivery usually produces a more manageable reading rhythm.
Why are results split into Latest Research, Most Relevant, and Classics?
The three sections represent different reasons to read a paper. Latest Research emphasizes recency, Most Relevant emphasizes fit to the alert context, and Classics helps recover established work that may frame the topic even when it is not new.
Do research alerts replace a systematic-review search?
No. Alerts support ongoing discovery, but they do not provide a complete, reproducible systematic-review protocol across every required database. Use the preview and stored recommendations as a monitoring layer alongside the formal methods your review requires.
Related guides
Connect monitoring to the rest of the workflow
An alert becomes more useful when its sources and results have somewhere durable to go.
Turn a recurring search into a reviewable alert
Define the source context, inspect a live preview, and choose a schedule that fits the topic.