AI Agent Frameworks and Runtimes
- earendil-works/pi: Unified LLM API, agent loop, TUI, and coding agent CLI in one toolkit (+401)
- Panniantong/Agent-Reach: Gives agents scraping access to Twitter, Reddit, YouTube, GitHub, and more via one CLI with no API fees (+980)
- Nasiko-Labs/nasiko: Open runtime for AI agents, written in Rust (+5)
- tinyhumansai/openhuman: Fast, low-cost open-source agent harness written in Rust (+98)
- ifixai-ai/iFixAi: Independent auditing layer for AI agents, answers whether an agent is doing what it is supposed to do in under 120 seconds (+298)
- experientiallabs/experiential: Zero-markup LLM gateway supporting BYOK, self-hosted models, and 1000-plus marketplace models with traffic-based cost optimization (+472)
Coding Assistants and Context Management
- anthropics/claude-code: Agentic coding tool that lives in your terminal and handles routine tasks, explanations, and git workflows via natural language (+337)
- thedotmack/claude-mem: Persistent cross-session memory for coding agents, compresses and reinjects context for Claude Code, Codex, Gemini, Copilot, and others (+628)
- mksglu/context-mode: Context window optimizer for AI coding agents, sandboxes tool output with a 98% reduction and routes across 17 platforms via MCP (+190)
- garrytan/gstack: Opinionated Claude Code setup with 23 tools covering CEO, designer, eng manager, release manager, doc engineer, and QA roles (+125)
- zeronsh/zeron: Native control plane for Claude Code, Codex, Cursor, Devin, and other coding agents, written in Rust (+81)
- AlexsJones/llmfit: Command-line tool that finds which models from hundreds of providers actually run on your local hardware (+64)
- pingdotgg/t3code: A TypeScript project from the t3 ecosystem, inferred to be a code tooling utility (+490)
Infrastructure and Platform Tooling
- cloudflare/cloudflare-os: Agent workspace on Cloudflare Workers for building apps, creating documents, and running agents with company context (+335)
- FalkorDB/FalkorDB: Graph database using GraphBLAS sparse adjacency matrices, targeted at GraphRAG and LLM knowledge graphs (+164)
- Effect-TS/effect: TypeScript library for building production-ready applications with typed effects and error handling (+243)
- getsentry/sentry: Developer-first error tracking and performance monitoring platform (+152)
- zerx-lab/FluxDown: Multi-protocol download manager in Rust supporting HTTP, FTP, BitTorrent, HLS, and DASH with multi-threaded acceleration (+299)
- p-e-w/heretic: Automatic censorship removal layer for language models (+171)
Video, Media, and Creative Tooling
- heygen-com/hyperframes: Write HTML and render it as video, designed for agent-driven video production (+376)
- calesthio/OpenMontage: Open-source agentic video production system with 12 pipelines, 100-plus tools, and 700-plus agent skill files (+245)
- OpenCut-app/OpenCut: Open-source alternative to CapCut for video editing (+512)
- earthtojake/text-to-cad: Python toolkit giving AI agents the ability to generate and manipulate CAD models (+83)
Dev Workflow and Productivity
- tester-army/e2e: Next-generation end-to-end testing framework for web and mobile apps (+345)
- Bennettxai/FounderOS-DEMO: Open-source single-operator business command center covering comms, funnel, social, finances, and agents from one dashboard (+10)
- mrifqidaffaaditya/WA-AKG: Self-hosted WhatsApp gateway and dashboard built with Next.js 15, Baileys, and Prisma (+31)
- hieunc229/mailflare: Email client built for professionals and teams (+241)
- jamwithai/production-agentic-rag-course: Course materials for building production agentic RAG systems (+220)
Finance and Data
- LuxAlgo/Vela: Fast, extensible open-source financial charting library for the web (+103)
- Mathieu2301/TradingView-API: TypeScript client for pulling real-time stock data from TradingView (+22)
- shy3130/tick-stock-panel: Self-hosted quantitative workbench for A-share stock screening, monitoring, and backtesting with LLM-driven strategy customization (+51)
- nkaz001/hftbacktest: High-frequency trading and market-making backtesting tool using full tick data with Level-2 and Level-3 order book support (+12)
The throughline across today's trending list is clear: builders are racing to extend, audit, and orchestrate AI coding agents rather than build new ones from scratch. The heavy clustering around context management, agent runtimes, and control planes suggests the ecosystem is entering a maturation phase where reliability and observability matter as much as raw capability.