See where open-source momentum starts and where it creates value
100,000 old stars tell you what was popular. +3,000 stars this week tells you where the world is moving now.
A focus-shattering terminal dashboard that turns your task list into a chaotic arcade game.
Inside-out: Its unusual game-like approach to attention management is spreading through productivity communities and social demos.
Real-time observability overlay for multi-agent systems with replayable traces.
Inside-out: Teams are moving from single AI assistants to agent groups and urgently need a way to understand failures.
Generate architecture diagrams straight from your codebase, no annotations required.
Inside-out: Its latest release produces useful diagrams with almost no setup, replacing hours of manual documentation.
An open coding agent that runs locally and keeps your keys on your machine.
Inside-out: Demand for capable local AI has jumped as companies push back on subscription cost and data privacy risk.
Convert PDFs, Office docs and images into clean Markdown for LLM pipelines.
Inside-out: AI search and document automation teams need a dependable way to prepare messy business files at scale.
A curated playbook of prompts, policies and eval suites for production agents.
Inside-out: High-profile agent mistakes have made reliable safety testing a board-level requirement rather than an optional extra.
Shared team workflows for AI coding sessions, right from the command line.
Inside-out: Teams want the speed of personal AI tools without every developer inventing a different process.
One fast Rust core that ships the same product across mobile, desktop, web, and the command line.
Inside-out: Teams are using it to avoid maintaining separate codebases for iPhone, Android, desktop, and web products.
Node-based canvas for designing and running data and model pipelines.
Inside-out: A major usability release made complex AI workflows understandable to people who do not want to write everything by hand.
Run private language models directly on iPhone, iPad, and Mac using Apple silicon.
Inside-out: New Apple silicon performance gains made useful private, offline AI practical on everyday phones and laptops.
A tiny, fast crawler tuned for building fresh retrieval indexes.
Inside-out: It recently showed faster fresh-content indexing with much lower infrastructure needs than larger crawling stacks.
A Jetpack Compose kit for generating adaptive Android interfaces from structured prompts.
Inside-out: A viral demo showed complete adaptive Android screens appearing in seconds while remaining native and accessible.
High-throughput, memory-efficient inference and serving engine for LLMs.
Inside-out: New model launches and benchmark gains keep making it the default performance layer for self-hosted AI services.
A high-performance Flutter canvas for responsive vector animation across mobile, desktop, and web.
Inside-out: Its new renderer matched native animation speed while keeping a single design and codebase across devices.
Detect schema and embedding drift in production RAG stacks before users do.
Inside-out: Companies are discovering that AI quality can fall silently after launch, creating demand for continuous monitoring.
An embedded database that ships with sane defaults and zero knobs to turn.
Inside-out: Builders are choosing simpler local-first tools over costly hosted databases for small and offline applications.
Streaming ETL with typed pipelines and one-command local replay.
Inside-out: Its typed workflow update makes stream processing safer and easier to test without a large operations team.
Accessible component library with motion baked in and zero runtime CSS.
Inside-out: Teams are adopting it after accessibility improvements and strong performance comparisons against larger UI kits.
Audit, repair and sync your shell setup across every machine you own.
Inside-out: A viral demo showed it fixing common developer setup problems automatically instead of requiring manual scripts.
Train small models on a single GPU with a two-file training loop.
Inside-out: Recent results showed competitive models trained without a large cloud cluster, attracting budget-conscious teams.
Parallel test runner and build cache for large Rust workspaces.
Inside-out: Benchmark results from large codebases showed dramatic build-time savings and triggered word-of-mouth adoption.
Rotate and distribute secrets to edge runtimes without a control plane.
Inside-out: Growth in edge applications has exposed how difficult secret rotation is across many providers and regions.
Continuous profiler that flags regressions in pull requests automatically.
Inside-out: Its pull-request integration makes performance testing automatic at a time when cloud efficiency is under scrutiny.
Browser-based pixel art studio with animation timelines and export presets.
Inside-out: Creators are sharing its instant browser workflow as a simpler alternative to paid desktop art tools.