Global

reproducible-research
12 projects tagged “reproducible-research” that gained the most stars yesterday
#896 · +37 · 1d ·GitHub

The open-source AI research workbench for scientific research and agent workflows. Local-first, model-agnostic desktop app with extensible skills, MCP tools and connectors, Python/R execution and traceable artifacts for reproducible research on macOS, Windows and Linux.

Creator
aipoch
Size
138.3 MB
Stars
5.3K
Age
3mo
Commits · 90 Days
#1646 · +17 · 1d ·GitHub

Token-scale scientific discovery engine — autonomous hypothesis generation, experiment execution, and knowledge graph synthesis

Size
85.1 MB
Stars
1K
Age
7mo
Commits · 90 Days

> ### 🚀 New here? Open the Skill Search → to filter all 1,096 skills by method, stage, language, and license. The 5-minute tour (make quickstart) prints the same picture in your terminal.

Size
42.3 MB
Stars
4.7K
Age
6mo
Commits · 90 Days
#2915 · +7 · 1d ·GitHub

The series is running. 5 days after launch. It runs once a day for 30 days: days 1–10 are the baseline, then there are two 10-day windows, so the first possible call is around 2026-10-24.

Creator
ninjahawk
Size
1.5 MB
Stars
1.5K
Updated
CommitOct 18
Age
27d
Commits · 90 Days
#3204 · +6 · 1d ·GitHub

Conditional integer multiplication: kappa > 2^-15. Community proofs, exact certificates and reproducible audits; assumes the OpenAI #109 framework.

Creator
CrocSwap
Size
1.9 GB
Stars
288
Age
13d
Commits · 90 Days
#5013 · +2 · 1d ·GitHub

Kedro is a toolbox for production-ready data science. It uses software engineering best practices to help you create data engineering and data science pipelines that are reproducible, maintainable, and modular.

Creator
kedro-org
Size
245.5 MB
Stars
11K
Age
7y
Commits · 90 Days
#5174 · +2 · 1d ·GitHub

Open memory & reliability layer for AI agents + independent judging/verification (NACRE judge, caliber-bench meta-benchmark, preregistered criteria) — by Nautilus Platform

Creator
chunxiaoxx
Size
425.5 MB
Stars
1.3K
Age
5mo
Commits · 90 Days
#8064 · +0 · 1d ·GitHub

Scientific hypothesis-discovery workbench: grounded hypothesis-evidence graphs, synthetic literature with planted causal chains, generate-debate-evolve loops over pluggable offline adapters, Elo tournaments recovering planted skill order, paraphrase dedup, Bayesian evidence…

Creator
OpSafari
Size
1.5 MB
Stars
555
Age
20d
Commits · 90 Days
#8280 · +0 · 1d ·GitHub

Without training models, RoboHarness combines visual clicks, keypoint tracking, and geometric constraints to provide an embodied control panel for an LLM agent.

Creator
BinceQu
Size
668.5 MB
Stars
249
Updated
CommitOct 16
Age
13d
Commits · 90 Days
#8497 · +0 · 1d ·GitHub

--- ## 🎯 Overview XRepoTest is a repository-level benchmark for evaluating how well large language models generate unit tests for real-world functions across 5 programming languages.

Creator
solis-team
Size
2.1 MB
Stars
167
Updated
CommitAug 27
Age
1mo
Commits · 90 Days
#8582 · +0 · 1d ·GitHub

Credit and method: Rajmoni created the original motion design. This project adapts that supplied reference: 180 text-cleared source frames preserve its backgrounds and control animations, while JavaScript renders the replacement ChatGPT typography.

Size
21.4 MB
Stars
140
Updated
CommitSep 06
Age
1mo
Commits · 90 Days