About this directory
Agentic Research Directory helps researchers explore documented uses of AI in academic work and assess what might be relevant to their own research.
How this directory is developed
The directory is developed by an experienced academic using an agentic AI approach. Specialist AI agents support resource discovery, source review, writing and technical maintenance, guided by an academic perspective on research needs and evidence quality.
This approach supports the directory’s development; it does not mean that every listed tool has been tested or every reported result independently reproduced.
Who it is for
The directory is intended for researchers, research students and educators across the humanities, social sciences, education, business and STEM. It covers task assistance as well as agentic and autonomous workflows: greater automation does not mean stronger evidence.
What you can find
Resources include tools, reusable agent skills, frameworks, workflows, tutorials, templates, documented use cases, and research-relevant datasets and benchmarks. They span activities such as literature research, data analysis, research coding, writing and experimental work.
Entries describe the documented approach, research relevance, adoption requirements and limitations, with links to supporting sources.
How to read the evidence
All entries are reviewed before publication and linked to sources. Review scope varies: it may cover a README, selected documentation, a paper excerpt or an abstract. Source checking is not independent reproduction, and inclusion does not establish a research benefit.
Source labels describe where an entry is grounded; outcome-verification labels indicate what has been checked about reported results. Read both alongside the entry's limitations and reproducibility notes. See Method for the distinctions.
Limits and responsible use
Coverage is selective, not exhaustive, and source checks are time-specific. Documentation, software and access requirements can change. Inclusion is not an endorsement or a guarantee of accuracy, safety, availability or effectiveness.
Before adoption, inspect the original sources, check licensing and data-handling requirements, and assess whether the methods and outputs are suitable for your research.
Feedback and corrections
To flag an error, an overstated claim or a reproduction attempt, use the comment form on the relevant resource entry. Include supporting sources where available. Comments are held for moderation before appearing.
Explore the directory or consult Updates for recorded discovery activity and reported source-access failures.