PullRepo

Daily radar for the fastest-growing AI tools & repos

Today's AI Research: Fastest-Growing Projects — April 15, 2026

Today's AI Research, we're seeing a surge of interest in tools that facilitate machine learning from first principles, autonomous improvement loops, and research agents. The top-growing repositories are focused on providing engineers with the ability to reason about ML systems and explore new ways of augmenting human thinking with AI.

Dreddnafious's `thereisnospoon` repository is gaining traction with a growth score of 41.79 and over 1,070 stars, as it offers a comprehensive machine learning primer built from first principles for engineers who want to reason about ML systems the way they reason about software systems. With 14 commits in the past 30 days, this project is actively being developed and refined.

Mskayyali's `nodepad` repository has a growth score of 38.31 and 790 stars, as it provides a spatial research tool that explores using AI to augment thinking, not replace it. With an impressive 76 commits in the past 30 days, this project is rapidly evolving and attracting attention from researchers.

Alvinreal's `awesome-autoresearch` repository boasts a growth score of 35.40 and over 1,283 stars, as it curates a list of autonomous improvement loops, research agents, and autoresearch-style systems inspired by Karpathy's autoresearch. With 41 commits in the past 30 days, this project is continuously updated with new resources.

WecoAI's `awesome-autoresearch` repository has a growth score of 24.56 and 867 stars, offering a curated list of AutoResearch use cases with optimization traces and open-source implementations. Although it has fewer commits than other projects on this list, its steady growth suggests ongoing interest in autoresearch applications.

Toby-bridges' `api-relay-audit` repository is gaining attention with a growth score of 19.41 and 219 stars, as it provides a security audit tool for third-party AI API relay/proxy services that detects hidden prompt injection, prompt leakage, instruction override, and context truncation. With 54 commits in the past 30 days, this project is actively addressing pressing security concerns.

Thunlp's `OPD` repository has a growth score of 16.50 and 33 stars, as it serves as the official repository for the paper "Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe". Although it has fewer stars than other projects on this list, its focused scope and recent activity suggest a dedicated following.

Fagemx's `gstack-game` repository boasts a growth score of 13.21 and 25 stars, offering a complete game production workflow for Claude Code that includes design review, prototype planning, and QA with fix loop. With an impressive 100 commits in the past 30 days, this project is rapidly developing and showcasing its versatility.

X-zheng16's `Awesome-Embodied-AI-Safety` repository has a growth score of 7.24 and 65 stars, providing a comprehensive survey of risks, attacks, and defenses in embodied AI. With 35 commits in the past 30 days, this project is continuously updated with new research.

SYuan03's `Skill-Anything` repository has a growth score of 6.12 and 221 stars, offering a tool that converts any source into an interactive learning package with quizzes, flashcards, and spaced repetition. Although it has fewer commits than other projects on this list, its steady growth suggests ongoing interest in AI-powered learning tools.

Icip-cas's `OmniBehavior` repository rounds out our list with a growth score of 5.67 and 32 stars, focusing on real-world human behavior simulation and benchmarking large language models on long-horizon, cross-scenario, heterogeneous behavior traces. With recent activity and a focused scope, this project is worth keeping an eye on.
Back to all reports