Free research prompt tool

Claude Research Prompt Library

Find a prompt by research stage, output type, and source discipline. Each prompt includes adaptation notes that keep Claude tied to evidence instead of inventing citations.

claude prompts for researchresearch prompt libraryclaude literature review prompts

Interactive workspace

StageTaskSearch prompts

Showing all prompts.

Search design / Literature review

Database search term expansion

I am reviewing [topic] in [population/corpus] from [date range]. Generate candidate search terms, controlled-vocabulary candidates, synonyms, adjacent terms, and likely false positives. Return a table with term, why it matters, where to test it, and what it might accidentally retrieve. Do not claim these searches have been run.

Adaptation note: Use before searching PubMed, OpenAlex, arXiv, or field databases. Save the final search string separately.

Screening / Literature review

Title and abstract screening

Classify each exported record against these inclusion criteria: [criteria]. Use labels include, exclude, maybe, and wrong corpus. Give one abstract-grounded reason and one short phrase from the abstract as the anchor. If the abstract is insufficient, label maybe.

Adaptation note: Keep the model output as a triage queue. A human reviewer owns the final decision.

Paper reading / Summarization

Fielded paper summary

Read the supplied paper text as a methodological reviewer. Return fields for research question, study design, sample or corpus, method, main claim, strongest evidence, stated limitations, unstated limitations, and follow-up searches. Tie every field to a passage, section, figure, or table marker.

Adaptation note: Use for one paper at a time. Avoid freeform summaries when a note needs to be auditable.

Paper reading / Methods critique

Methods and measurement critique

Assess the methods section only. Identify the study design, variables, measures, sampling frame, statistical or computational approach, missing controls, and assumptions. Separate what the authors report from your critique. Mark any field as not reported when absent.

Adaptation note: Run after the fielded summary so methods concerns do not get buried in prose.

Extraction / Evidence table

Evidence table extraction

Extract the following fields only from the provided source: citation, study design, setting, sample or corpus, method, intervention or exposure, comparator, outcome, effect or finding, limitation, funding or conflict, and passage marker. Use not reported when absent. Do not infer missing values from background knowledge.

Adaptation note: Paste the output into a spreadsheet and verify important fields against the source.

Synthesis / Gap analysis

Evidence gap map

Using only the supplied evidence table, identify gaps by population, method, outcome, data source, time period, and evidence quality. For each gap, list the source IDs that make the gap visible and one search query that could test whether newer evidence exists.

Adaptation note: Useful for dissertation background, grant significance sections, and review discussion drafts.

Drafting / Grant drafting

Grant background from locked evidence

Draft a grant-background paragraph from the supplied evidence table only. Use cautious verbs, cite source IDs after each sentence, and add a final bullet list called "claims needing stronger evidence." Do not add papers or statistics that are not in the table.

Adaptation note: Designed for background drafts, not for final submission without expert and source review.

Verification / Contradiction review

Skeptic pass for synthesis

Review this synthesis against the evidence table. For each paragraph, identify the strongest source that complicates it, any unsupported verb, any missing qualifier, and a safer rewrite. If no source supports a sentence, mark it unsupported.

Adaptation note: Run before sending a literature review section to a supervisor, PI, or reviewer.

Verification / Citation checking

Citation support audit

For each cited sentence, return the citation, exact claim, source passage or marker, whether the passage supports the verb strength, and the evidence level: full text, abstract only, metadata only, or unverified. Flag any citation that appears real but does not support the sentence.

Adaptation note: Pair with the citation integrity checker when Claude or another model generated citations.

Notebook / Reproducibility

Computational reproducibility checklist

Create a reproducibility checklist from this method and artifact history. Include data availability, code availability, environment, package versions, random seeds, hardware or compute assumptions, external tools, and any step that cannot be reproduced from the record.

Adaptation note: Use for computational papers, Claude Science artifacts, and code-backed research notes.

Notebook / Lab notebook

AI-use notebook entry

Turn this AI-assisted research step into a notebook entry. Include date, tool or product surface, prompt purpose, input sources, output summary, human decision, corrections made, unresolved checks, and links to saved artifacts. Keep it concise and auditable.

Adaptation note: Use immediately after a meaningful AI-assisted step, while source state is fresh.

Library / Reference management

Zotero library cleanup plan

Review this exported reference list. Identify duplicate records, missing identifiers, unclear source types, records needing DOI or PMID checks, and collections or tags that would make the library easier to audit. Do not invent missing metadata.

Adaptation note: Paste exported metadata, not private library credentials.

Ready.

Privacy note: this tool is client-side only. Inputs stay in your browser unless you copy, download, or paste the output somewhere else.

How to use this tool

Choose the research stage or leave it set to all prompts.
Filter by task or search for a phrase.
Open a prompt and read the adaptation notes.
Copy the prompt and replace bracketed fields with project details.
Save the final prompt and Claude output in your research notebook.

FAQ

Are these prompts enough to prevent hallucinated citations?

No prompt can guarantee that. The prompts reduce risk by requiring source anchors, missing-value labels, and verification steps.

Can I use these prompts with non-Claude tools?

Usually yes. They are written for Claude Researcher workflows, but the evidence rules apply to any AI assistant used in scholarly work.

Related workflows

Prompt templatesPromptsPrompt templates for literature search design, paper triage, evidence extraction, contradiction review, and reproducibility checks.Literature review protocolLit ReviewA source-first workflow for using Claude in a literature review without losing PRISMA discipline, search provenance, or human judgment.Audit trailNotebookHow to preserve prompts, search strings, source exports, Claude outputs, code, figures, and verification notes as an auditable research record.

More free tools

Protocol BuilderBuild an exportable literature review protocol and Claude prompts for search design, screening, extraction, and synthesis.Citation CheckerScore the risk of an AI-assisted citation by checking existence, identifier resolution, claim support, retraction status, and evidence level.