Dataset & Code Finder
Dataset search for your question or papers: a ranked table of datasets, code repositories and configs, with coverage, access, license, quality and activity notes.
Built for researchers, analysts and engineers who run dataset searches to answer a question or reproduce work.
- Research news verification
Why this work is hard
Researchers, analysts and engineers who need data or code to answer a question or reproduce work spend hours on this by hand: finding the right sources, reading them, checking them against each other and writing up what matters. The work is repetitive, it is easy to miss something, and nobody can show where a claim came from.
Dataset & Code Finder does that work as a GREP AI agent. It takes a whole list in one run and returns one row per item. Every run delivers a sourced report, with every claim linked to its source.
What you get
How it works
- 01
Fill in the form
The form asks for: Question, topic or papers (one per line); Resource types wanted (optional); Constraints: license, format, region, period (optional).
- 02
Skills run
1 specialized skill runs across live sources, registries and documents: research news verification.
- 03
Evidence and checks
Every claim is cross-referenced and linked to the source it came from.
- 04
You get the result
Sourced report, ready to share, export or run again.
What Dataset & Code Finder delivers
FAQ
What does Dataset & Code Finder need to run?
Dataset & Code Finder needs question, topic or papers (one per line); resource types wanted (optional) and constraints: license, format, region, period (optional) are optional.
How often does Dataset & Code Finder run?
It takes a whole list in one run and returns one row per item.
What does Dataset & Code Finder deliver?
Every run delivers a sourced report, with every claim linked to its source.
Can I change what Dataset & Code Finder does?
Yes. Every GREP AI agent can be cloned into your workspace, where you can edit its instructions, add your own sources and files, and change its schedule.
Where does the dataset search look?
The paper's own links, then GitHub and GitLab, Hugging Face, Zenodo, Dryad and Figshare, government open-data portals, Kaggle and UCI, and domain archives. Commercial sources are named only when no open source fits.
How are licenses checked?
From the actual LICENSE file or dataset card, never from a summary. Each link is checked to resolve, and the last commit or update date is recorded.
What if a resource fails one of my constraints?
It is listed as a near miss that names the failed constraint, rather than dropped silently.
More agents like this
- Runs on a schedule, weekly
Citation Alert
Citation alerts for your seed papers or authors: each run lists new citing papers with abstract, venue, method, and exactly how they use the cited work.
Formatted briefView → - Runs on a schedule, weekly
Open Model & Dataset Release Watch
A weekly AI model release watch for your domain: new, updated or retired open models and datasets, with license, size, access and benchmark comparisons to what you use.
Formatted briefView → - Runs on a schedule, daily
Job Opening Watch
AI job alerts that run daily and give you a ranked, deduplicated list of new postings matching your role, location and criteria, with pay, deadlines, links and fit notes, delivered as a spreadsheet.
SpreadsheetView → - Runs over a list
Citation Verifier
A citation checker for every reference: existence, accuracy, retraction or overruling, and misquotation, with a keep, fix or drop verdict and a corrected reference.
Sourced reportView → - Runs on a schedule, monthly
Data Series and Price Index Refresh
Price index tracking for your named data series and price indices: a refreshed table and charts, with moves against prior periods and what they mean for your budget or plan.
Formatted briefView → - Runs on demand
Funding Finder
A grant search that returns a ranked list of open grants, scholarships, fellowships, tax credits and incentives you qualify for, with amounts, eligibility, deadlines and application links.
SpreadsheetView →
Run Dataset & Code Finder in minutes.
Free to start. Run it once, put it on a schedule, or clone it and make it yours.