Use AI to write your PubMed query.

The workflow

Let AI draft the logic, not hide the search.

1. Write the real question

   Describe the population, problem, intervention or exposure, and outcome when those elements matter. Plain language is fine.

2. Build the search blocks

   Ask AI for the major concepts, possible MeSH terms, common synonyms, spelling variants, and a broad and focused version of the Boolean query.

3. Run it in PubMed

   Paste the query into PubMed. Inspect the first results and Search Details, confirm any MeSH terms, and widen or narrow the query based on what you actually retrieve.

I want to search PubMed for this clinical question:

[Write your question here. Include the population, problem, intervention or exposure, comparison, and outcome when relevant.]

Help me build a transparent PubMed search strategy.

1. Identify the main concepts in my question.
2. For each concept, suggest relevant MeSH terms, common synonyms, spelling variations, abbreviations, and related free-text terms.
3. Combine synonyms within each concept using OR.
4. Combine the major concepts using AND.
5. Give me two copyable PubMed queries:
   - a broad, sensitive version
   - a more focused version
6. Explain what each search block is doing.

Do not invent MeSH terms. Clearly label any term that I should confirm in the PubMed MeSH Database. Do not summarize papers or claim that the search is comprehensive.

Before you trust the search

Inspect what PubMed actually did.