Bookmarks
Trending Movers — Full Archive
AI/ML
- PrimeIntellect-ai/prime-agent
- Open-source coding and research agent for general, long-running work, built around a Recursive Language Model (RLM) that treats context as variables. It's part of the push toward self-improving RL agents for autonomous coding workflows.
- TencentCloud/TencentDB-Agent-Memory
- Team-level memory hub for AI agents that turns conversations, docs, and code into four reusable memory assets — Chat Memory, Skill, LLM-Wiki, and Code-Graph — governed and shared across agent frameworks.
- vitali87/code-graph-rag
- Parses multi-language monorepos with Tree-sitter, builds a knowledge graph in Memgraph, and lets you query, understand, and edit the codebase in plain English.
- huangruiteng/loopx
- Provider-neutral stateful control plane for long-running agents — durable goals, quota-aware auto-wake, executable todos, evidence logs, and verifiable handoffs across Codex, Claude Code, and other agent loops.
- volcengine/OpenViking
- An open-source context database for AI agents that unifies agent memory, knowledge RAG, and skills. It stores memories and resources as queryable context, with docs, a live demo, and community channels linked from the repo.
- cactus-compute/needle
- Needle 2 is an open 45M-parameter model for tool calling, device use, and structured extraction, compressed into a single 14MB binary (CQ2-bit) that runs a full session in roughly 28MB of RAM. It targets phones, wearables, smart home devices, and robots.
- modular/modular
- The repository for open-source components of the Modular Platform, including the MAX framework for model serving and the Mojo language. It positions Modular as a unified platform for AI development and deployment.
- harry0703/MoneyPrinterTurbo
- A one-stop AI short-video generator: given a topic or keyword, it automatically produces a script, matches stock footage, generates subtitles and background music, and composites an HD video via an automated workflow. The project ships a WebUI and API.
- lightningpixel/modly
- A desktop app for Windows, Linux, and Apple Silicon that generates 3D meshes from images or text prompts using open-source AI models running entirely on the local GPU.
- unslothai/unsloth
- Unsloth's first desktop app for running and training LLMs and diffusion models locally. It covers models including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, and FLUX.
- freestylefly/awesome-gpt-image-2
- A Chinese-language prompt library for GPT-Image-2, built from more than 530 reverse-engineered examples and organized into over 20 reusable templates. It's a snapshot of how quickly prompt-engineering know-how for one model can turn into a shared reference once enough people start experimenting with it at the same time.
- santifer/career-ops
- An open-source job-search agent that scans listings, scores each one A through H, tailors a resume to match, and tracks applications, run locally from inside a coding CLI like Claude Code or Codex. Both its web and core packages shipped new releases on August 25, the same day the star count spiked, which points to the release itself as the trigger.
- chaitanyagiri/munder-difflin
- An Electron desktop app that turns several coding-agent CLIs — Claude Code, Gemini, Codex, Grok — into a coordinated team, complete with persistent memory, an inter-agent mailbox, and a "GOD agent" that routes tasks and only escalates critical decisions to a human. It shipped v0.4.5 on August 22 with fixes to cost reporting and Apple Silicon memory embeddings, so the surge lines up with active, visible development rather than a single headline feature.
- xbtlin/ai-berkshire
- An AI-era value investing research framework built for Claude Code/Codex, combining four masters' methodologies with multi-agent adversarial analysis.
- DeusData/codebase-memory-mcp
- A high-performance code intelligence MCP server that indexes codebases into a persistent knowledge graph in milliseconds, supporting 158 languages with sub-ms queries.
- calesthio/OpenMontage
- The world's first open-source, agentic video production system with 12 pipelines, 52 tools, and 500+ agent skills, turning AI coding assistants into full video production studios.
- Robbyant/lingbot-map
- A feed-forward 3D foundation model for reconstructing scenes from streaming data, focused on geometric context transformation.
- topoteretes/cognee
- An open-source AI memory platform for agents, giving AI agents persistent long-term memory across sessions with a self-hosted knowledge graph engine.
- msitarzewski/agency-agents
- A complete AI agency at your fingertips, with specialized expert agents for frontend, Reddit community management, whimsy injection, and reality checking.
- interviewstreet/hiring-agent
- An AI agent that evaluates and scores resumes, extracting structured data from PDFs, enriching with GitHub signals, and outputting fair, explainable evaluations.
- MadsLorentzen/ai-job-search
- An AI-powered job application framework built on Claude Code — fork it, fill in your profile, and let Claude evaluate jobs, tailor CVs, write cover letters, and prepare you for interviews.
- Nutlope/hallmark
- Anti-AI-slop design skill for Claude Code, Cursor, and Codex that refuses to look AI-generated, picking macrostructure from twenty themes and running fifty-seven slop-test gates plus a pre-emit self-critique.
- HKUDS/DeepTutor
- Lifelong personalized tutoring system with CLI, ecosystem, and community features, recently releasing v1.5.3 with new tutor capabilities.
- PrismML-Eng/Bonsai-demo
- Demo repository for running Bonsai language models locally on Mac (Metal), Linux/Windows (CUDA, Vulkan, ROCm), or CPU, including the new Bonsai 27B model.
- diegosouzapw/OmniRoute
- Free MIT AI gateway providing one endpoint for 290+ providers (90+ free) and 500+ models, with quota-aware auto-fallback, RTK+Caveman compression saving 15-95% tokens, and MCP/A2A support.
- Shubhamsaboo/awesome-llm-apps
- 100+ open-source AI agents, agent skills, and RAG apps — hand-built, tested end-to-end, Apache-2.0 licensed, working with Claude, Gemini, GPT, DeepSeek, Llama, Qwen and other open-source models.
- asgeirtj/system_prompts_leaks
- Extracted system prompts from Anthropic, OpenAI, Google, xAI, and more — updated regularly and featured in The Washington Post.
- google-research/timesfm
- TimesFM (Time Series Foundation Model) is a pretrained time-series foundation model from Google Research for forecasting.
- jamiepine/voicebox
- The open-source AI voice studio for cloning voices, generating speech, and dictating into any app — running locally.
- Panniantong/Agent-Reach
- Give your AI agent eyes to see the entire internet — read and search Twitter, Reddit, YouTube, GitHub, and more with one CLI and zero API fees.
- withastro/flue
- The sandbox agent framework — build autonomous agents and powerful AI workflows with programmable TypeScript harness.
- debpalash/VoiceStudio
- Its pitch is a fully local, open alternative to ElevenLabs: voice cloning, voice design, video dubbing, dictation, and audiobook narration across 646 languages, with nothing sent to a cloud API.
- JustVugg/colibri
- Written in plain C with zero dependencies, colibri streams a mixture-of-experts model's experts off disk rather than holding the whole thing in memory, which is how it gets a frontier-scale model running on hardware that couldn't otherwise fit it.
- THU-MAIC/OpenMAIC
- Tsinghua's multi-agent interactive classroom spins up a set of cooperating agents to run an immersive lesson from a single click, built for educators rather than developers.
- TauricResearch/TradingAgents
- Each agent in TauricResearch's framework plays a distinct role — analyst, risk manager, trader — and the group has to agree before a position gets opened, modeling a trading desk rather than a single autonomous bot.
Data
- semantica-agi/semantica
- Graph-native infrastructure that ingests enterprise data, builds a knowledge/context graph, and runs graph analytics and causal reasoning with decision provenance baked in. Positioning itself as an open-source "Palantir for AI agents" for explainable, accountable systems.
- apache/ossie
- Apache Ossie is an industry-wide specification effort to standardize how we exchange semantic metadata across analytics, AI and BI platforms, providing a vendor-neutral single source of truth for semantic data.
- ZhuLinsen/daily_stock_analysis
- LLM-powered multi-market stock analysis system with multi-source data, real-time news, decision dashboard, and automated notifications.
Dev Tools
- google/skills
- Agent Skills for Google products and Google Cloud, installable with `npx skills add google/skills`. Riding the wave of skill-pack standardization for AI coding agents.
- addyosmani/agent-skills
- Production-grade engineering "skills" that encode senior-engineer workflows and quality gates so AI coding agents follow them consistently across define, plan, build, verify, review, and ship phases.
- firecrawl/pdf-inspector
- Fast Rust library for PDF inspection and classification: detects scanned vs. text-based PDFs, extracts position-aware text, and converts to clean Markdown without OCR. Ships Python bindings for smart routing decisions.
- esengine/DeepSeek-Reasonix
- DeepSeek-native AI coding agent for the terminal — a single Go binary engineered around prefix-cache stability, with plan mode, permissions, a workspace sandbox, and four surfaces: terminal, desktop, browser, and editor via ACP.
- cathrynlavery/diagram-design
- A collection of 38 editorial diagram types delivered as self-contained HTML + SVG, aimed at AI coding agents such as Claude Code, Codex, and Pi. Version 2.0 adds the Loop flywheel pattern with a shared-memory hub, and 2.3 adds semantic system patterns and optional accessible motion.
- akitaonrails/ai-memory
- A long-term memory solution for AI coding CLIs that lets you switch mid-task between agents such as Claude Code and OpenAI Codex in the same directory without re-explaining architecture or open questions. Linux is the primary supported target.
- openai/codex
- OpenAI's terminal-based coding agent pushed three alpha builds in a single day this week — 0.150.0-alpha.9 through -11 — a release cadence that roughly matches its star trajectory. It's the most direct rival to Claude Code and Cursor in the CLI coding-agent space, and right now the shipping pace looks like as much of a story as any individual feature.
- Alishahryar1/free-claude-code
- A pass-through that runs Claude Code, Codex, Pi, and OpenCode against a pool the author says totals over 1.3 billion free tokens, reachable from a terminal, an app, an IDE, or a phone. The pitch is entirely about the cost of running these coding agents rather than their capability, so it makes sense that it's spreading fastest among people who already use the tools it wraps.
- cordiverse/cordis
- A TypeScript meta-framework built around what its maintainers call "spatiotemporal composability," backed by a peer-reviewed paper on the underlying programming model. The API is explicitly marked unstable, which makes a 2,000-star week for a framework still in flux the more notable part of its story.
- ripienaar/free-for-dev
- A comprehensive list of SaaS, PaaS, and IaaS offerings with free tiers of interest to devops and infrastructure developers.
- kunchenguid/no-mistakes
- A local git proxy that runs an AI-driven validation pipeline before pushing, spinning up disposable worktrees and opening clean PRs automatically.
- google-labs-code/design.md
- A format specification for describing visual identity to coding agents, giving agents a persistent, structured understanding of a design system.
- JCodesMore/ai-website-cloner-template
- A reusable template for reverse-engineering any website into a clean, modern Next.js codebase using AI coding agents with a single command.
- stablyai/orca
- An AI orchestrator for running multiple coding agents side-by-side — each in its own worktree, tracked in one place, available on desktop and mobile.
- alirezarezvani/claude-skills
- A massive collection of 345 Claude Code skills, agent skills, and plugins covering engineering, marketing, product, compliance, research, and daily productivity — compatible with Claude Code, Codex, Gemini CLI, Cursor, and 8 more coding agents.
- tirth8205/code-review-graph
- Local-first code intelligence graph for MCP and CLI that builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.
- 1jehuang/jcode
- The next generation coding agent harness designed for multi-session workflows, infinite customizability, and performance, raising the skill ceiling for AI-assisted development.
- ibelick/ui-skills
- Skills for Design Engineers — a CLI tool that routes your agent through the right UI skill set for the task, with categories for motion, baseline UI, and more.
- agegr/pi-web
- Local web UI for the pi coding agent that reads local pi session files and provides a browser workspace for session browsing, real-time chat, model configuration, skill management, and project file preview.
- earendil-works/pi
- AI agent toolkit with unified LLM API, agent loop, TUI, and coding agent CLI, including the @earendil-works/pi-coding-agent interactive coding agent.
- Stirling-Tools/Stirling-PDF
- ayghri/i-have-adhd
- A skill for coding agents that forces the answer to the top of the reply instead of burying it under paragraphs of exploratory reasoning first. It topped the week by a wide margin: the +17,658 gain is more than a third of the repo's entire 45,898-star history, arriving inside seven days.
- tt-a1i/archify
- Built as an agent skill rather than a standalone app, archify turns a codebase into architecture, sequence, and data-flow diagrams rendered as animated, self-contained HTML instead of static images.
- DietrichGebert/ponytail
- Already sitting at 139,000 stars before this week started, ponytail added another 7,722 — an unusually large jump for a repo this established. The premise is a coding-agent prompt built around restraint: the best line of code is the one never written.
- affaan-m/ECC
- ECC layers skills, memory, and security scaffolding onto Claude Code, Codex, Cursor, and other agent harnesses rather than replacing any of them. The project's total is the largest in this week's list, just under 259,000 stars, and it added 6,085 more in the past seven days.
- heygen-com/hyperframes
- hyperframes takes HTML as input and renders it out as video — built by HeyGen, aimed at agents that can write markup easily but have no native way to produce a video file.
- microsoft/markitdown
- Office documents, PDFs, and Word files go in; Markdown comes out. It's Microsoft's own converter for feeding real-world documents into an LLM pipeline that only wants plain text.
- obra/superpowers
- obra pairs a software-development methodology with the agent skills to carry it out, and the combination is already the single most-starred repo on this week's list — 286,948 stars total, plus 4,124 more this week.
- blader/humanizer
- It strips the same tells — inflated significance, rule-of-three lists, em-dash overuse — that this digest was itself edited to avoid, packaged as a skill an agent can run on its own output before publishing.
Infra/DevOps
- cloudflare/computer
- Virtual filesystem inside a Cloudflare Durable Object that gives agents a sandboxed computer, with authoritative state in SQLite and pluggable runtimes (container and FUSE-mount backends today). A notable infrastructure play for secure agent execution.
- NVIDIA-NeMo/Switchyard
- A Rust proxy and library for LLM traffic that routes requests across models and providers while preserving native OpenAI and Anthropic API compatibility. It translates between API formats and records operational metrics for benchmarking and cost/performance optimization.
Web
- drawdb-io/drawdb
- Free, browser-based database ERD editor and SQL generator — build diagrams with a few clicks and export/import SQL. One of the few non-agent projects in this week's top tier.
- LadybirdBrowser/ladybird
- A truly independent, pre-alpha web browser built on a novel standards-based engine with a multi-process architecture. The clearest non-AI breakout in this week's feed.
- saadeghi/daisyui
- The most popular free and open-source Tailwind CSS component library, providing a clean, customizable set of UI components for rapid web development.
- penpot/penpot
- The open-source design platform for teams that build digital products at scale, giving full ownership of design infrastructure.
- bilawalsidhu/gods-eye-view
- Real, publicly available satellite and open-source intelligence feeds get projected onto a photorealistic 3D globe you can fly around in a browser tab. The "spy satellite simulator" description undersells it a little — none of the imagery is fictional.
Other
- 3b1b/manim
- Precise programmatic animation engine for explanatory math videos, started by 3Blue1Brown's Grant Sanderson. The original repository behind the channel's visuals; the community-maintained fork lives elsewhere.
- TapXWorld/ChinaTextbook
- Open collection of Chinese K-12 and university PDF textbooks, gathered to make free educational resources accessible and counter reselling of watermarked copies.
- basecamp/omarchy
- An opinionated Linux distribution by DHH, positioned as beautiful and modern. The repo hosts the authoritative manual (mirrored to learn.omacom.io), including getting-started guides for users coming from Mac or Windows.
- public-apis/public-apis
- The long-running, community-maintained collective list of free APIs for developers and prototyping, continuously updated across hundreds of categories.
- AprilNEA/OpenLogi
- A native, local-first alternative to Logitech Options+ written in Rust: remap buttons, DPI, and SmartShift over HID++ with no account and no telemetry. It's under active development and not yet stable.
- macro-inc/macro
- An all-in-one team workspace that unifies email, chat, docs, tasks, agents, calls, and CRM in a single interface, with @-linked items and shared team-level AI memory.
- byoungd/up
- Not software at all: a long, Chinese-language guide to lifelong learning in the AI era, moving from English study through project-based work to personal reflection, built around the line "AI is making answers cheap, but what remains scarce is knowing which questions matter." Its presence here is the interesting part — a reminder that this leaderboard's weekly momentum isn't limited to code.
- mauriceboe/TREK
- A self-hosted, real-time collaborative travel planner with interactive maps, budgets, packing lists, journal, and AI features.
- koala73/worldmonitor
- Real-time global intelligence dashboard with AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface.
- OpenCut-app/OpenCut
- The open-source CapCut alternative — a free video editor for web, desktop, and mobile being rewritten from the ground up.
Security
- zhaoxuya520/reverse-skill
- Cybersecurity "skill router" pack for reverse engineering, authorized penetration testing, and security research. AI-powered routing plus on-demand toolchain bootstrapping for Claude Code, Kiro, Cursor, and Cline.
- megadose/holehe
- An OSINT tool that checks whether an email address is registered on sites like Twitter and Instagram by probing forgotten-password flows. It returns per-site registration status and associated metadata.
- mukul975/Anthropic-Cybersecurity-Skills
- A library of 817 structured cybersecurity skills for AI agents, mapped against MITRE ATT&CK, NIST CSF 2.0, MITRE ATLAS, D3FEND, the NIST AI RMF, and MITRE F3, built to the agentskills.io standard so it works across Claude Code, Copilot, Codex CLI, Cursor, and other agent platforms. Its last tagged release was back in June, so this week's jump reads more like discovery than a launch.
- simplex-chat/simplex-chat
- The first messaging network operating without user identifiers of any kind — 100% private by design, with iOS, Android, and desktop apps.
- usestrix/strix
- An open-source AI penetration testing tool that finds and fixes app vulnerabilities, integrating with GitHub Actions and CI/CD pipelines.
CLI
- MoonshotAI/kimi-code
- Kimi Code CLI is an AI coding agent that runs in your terminal — it can read and edit code, run shell commands, search files, fetch web pages, and choose the next step based on feedback, working out of the box with Moonshot AI's Kimi models.
- MoonshotAI/kimi-cli
- The predecessor to Kimi Code CLI — an AI agent that runs in the terminal, now being gradually wound down as it evolves into the next-generation Kimi Code CLI.