Unsloth is a full-stack, local-first AI platform for fine-tuning and inference of large language models (LLMs), diffusion models, and audio models. It consists of two primary components: (1) the core `unsloth` Python library, which provides highly optimized LoRA/QLoRA/RL training with custom Triton kernels, and (2) the Unsloth Studio application, a Tauri desktop app with a React frontend (TanStack Router, Zustand) and a FastAPI backend (SQLite, structlog). The platform enables users to discover models from the Hugging Face Hub, download and manage them locally, run inference with chat, image/video generation, and RAG, fine-tune with a visual recipe editor, and export in multiple formats (GGUF, NVFP4). It also integrates external LLM providers (OpenAI, Anthropic, Ollama) and supports agentic tool loops via MCP and Codex transports. The codebase is extensively tested with both unit tests and source-level contract tests, and includes cross-platform desktop installers for Windows, macOS, and Linux.
Mattermost is a comprehensive, self-hosted workplace collaboration platform designed for secure team communication. The codebase implements a full-stack architecture with a Go-based backend server handling business logic, API endpoints, and database interactions, alongside a React/TypeScript frontend for user interface and state management. It supports real-time messaging, file sharing, integrations, and enterprise features like SAML, LDAP, and content flagging. The system is designed for deployment as a single binary or via Docker/Kubernetes, with PostgreSQL as the primary data store. The architecture follows a monolithic server structure with a distinct webapp layer, utilizing Redux for frontend state management and interface-based design for backend services. Key entry points include the main Go server binary, the CLI tool `mmctl`, and the React application root. The platform emphasizes extensibility through a robust plugin system and maintains strict type safety across the frontend and backend via shared model definitions. It features extensive testing infrastructure including unit tests, E2E tests with Cypress and Playwright, and custom Go linters for code quality enforcement. The project demonstrates high engineering standards with active maintenance, evidenced by recent commits and a large contributor base. It manages technical debt through custom tooling and versioned migrations, though the complexity of the monorepo and dual testing frameworks presents ongoing maintenance challenges. The codebase is suitable for engineers seeking a mature, enterprise-grade collaboration solution or reusable patterns for full-stack application development.
ClawHub is a full-stack registry marketplace for AI agent skills and OpenClaw plugins. It provides a web application (React + TanStack Router + Convex backend), a CLI tool (clawhub) for publishing and managing skills, an admin CLI for moderation, and a public HTTP API. The system handles skill/plugin publishing, versioning, search (with OpenAI embeddings), trending leaderboards, security scanning (VirusTotal, ClawScan, static analysis), moderation, and catalog feeds. It also mirrors an external skills.sh catalog. The architecture is a monorepo with shared schema packages, extensive test coverage (unit, e2e, Convex), and CI/CD via GitHub Actions. While primarily an application, it includes data/research elements like security dataset export and ranking metrics pipelines.
This repository is a production-grade, AI-powered stock analysis platform for CN, HK, US, JP, KR, and TW markets. In production it runs a scheduled analysis pipeline that fetches market data from multiple providers (akshare, yfinance, tushare, efinance, futu, etc.) with priority-based fallback, feeds the data to LLM backends via LiteLLM, generates structured decision reports, and pushes them through 15+ notification channels (WeChat, Feishu, DingTalk, Telegram, email, Discord, Slack, etc.). It also exposes a FastAPI REST API, a React/TypeScript web dashboard, an Electron desktop client, and bot integrations for Feishu/DingTalk/Discord. The system includes agent-based chat with tool use, screening, backtesting, alerts, portfolio management, decision-signal tracking, and market review capabilities.
One API is a production-ready OpenAI API gateway that aggregates 30+ LLM providers (OpenAI, Anthropic, Google Gemini, Azure, Baidu, Alibaba, Zhipu, Moonshot, DeepSeek, Ollama, etc.) behind a single OpenAI-compatible endpoint. The Go backend implements a sophisticated relay system with provider-specific adaptors that translate OpenAI-format requests to each provider's native API. It features token-based quota management with group/channel multipliers, load balancing across channels, automatic failover, channel health monitoring with metric-based auto-disable, and multi-node deployment with Redis synchronization. The React frontend (three themes: default, berry, air) provides admin dashboards for channel/token/user management, billing visualization, and system configuration. The codebase demonstrates strong engineering practices: clean adaptor pattern for provider extensibility, comprehensive middleware chain, batched database updates for performance, and well-structured configuration management.
OpenClaw AWD Arena is a comprehensive platform for conducting Attack-With-Defense (AWD) cybersecurity competitions, specifically designed for training and evaluating large language models (LLMs) in offensive and defensive security scenarios. The system orchestrates isolated Docker container environments where each participant (LLM agent or human) runs a vulnerable service that must be defended while attempting to capture flags from opponents' services. It features a web-based interface for match configuration and real-time monitoring, a Python/FastAPI backend handling game logic, scoring, and WebSocket-based event streaming, and custom orchestration logic for managing the lifecycle of match environments. The platform includes built-in flag management, scoring systems, match replay functionality, and support for multiple agent backends via a plugin architecture.
Langflow is a full-stack, open-source platform for building and deploying AI-powered agents and workflows. It provides a visual flow builder (React frontend), a Python backend (FastAPI/SQLModel), a CLI executor (LFX), an MCP server for AI tool integration, and a Python SDK. The codebase is a monorepo with multiple packages: the main `langflow` distribution, `langflow-base` (core backend), `lfx` (lightweight executor/CLI), `langflow-sdk` (API client), `langflow-stepflow` (step-based execution), and extension bundles (e.g., lfx-duckduckgo, lfx-arxiv, lfx-ibm, lfx-docling). The backend uses SQLModel with Alembic migrations, Celery for background tasks, FastAPI for REST endpoints, and a plugin/extension system for bundles. The frontend is a React/TypeScript app built with Vite, React Flow, Zustand, and Tailwind. The architecture supports multi-worker deployments, auto-login, JWT auth, rate limiting, MCP server integration, and a rich component library for LLMs, vector stores, tools, and data processing.
Packmind is a developer productivity platform that centralizes engineering standards and distributes them to AI coding agents. It solves the problem of scattered standards (in Slack, Notion, PR comments) that AI agents can't access and the manual copy-pasting of standards across repositories. The platform provides organization/space-based collaboration, Git provider integrations, LLM/AI service abstraction, marketplace for sharing components, deployment/target management, and change proposal workflows. It features a hexagonal architecture with clear separation of concerns across domains (accounts, organizations, spaces, standards, skills, commands, deployments), using NestJS for the backend, React for the frontend, and a separate CLI tool for local artifact management.
ListenBrainz Server is the backend and frontend for ListenBrainz, an open-source music listening history platform operated by the MetaBrainz Foundation. In production it ingests listening events from multiple sources (Spotify, Last.fm, Libre.fm, Maloja, file uploads, and direct API submissions), stores them in TimescaleDB, enriches them with MusicBrainz identifiers via an MBID mapping pipeline, and exposes them through a public REST API, a Last.fm-compatible API, WebSocket streams, and a React single-page application. The system also computes per-user and site-wide statistics, artist/recording/user similarity, collaborative-filtering recommendations, and annual 'Year in Music' summaries using Apache Spark batch jobs orchestrated through RabbitMQ and HDFS. The architecture is a hybrid of a traditional Flask web application and a set of background data-processing services. The main web server handles user authentication, API endpoints, playlist management, social features, and page rendering, while separate workers consume RabbitMQ queues for timescale writes, Spark job requests, metadata cache crawling (Spotify, Apple Music, SoundCloud, Internet Archive), and background imports/exports. The frontend is a large TypeScript/React application with React Router, TanStack Query, Jotai state, and a custom multi-source audio player called BrainzPlayer. Although the repository is primarily an application, it contains substantial data-engineering machinery—Spark incremental stats engines, dump generation/import, MBID canonicalization, and similarity algorithms—making it a genuinely hybrid codebase with the application character dominant. The codebase is well-structured for its size: it has clear module boundaries, abstract base classes for importers and OAuth services, a reusable incremental stats query-provider pattern, and extensive unit and integration tests. However, it also carries significant complexity and technical debt: multiple database backends, duplicated SQL across Spark entity stats, legacy compatibility endpoints, hardcoded external URLs, and some skipped or heavily mocked tests. The project is actively maintained, with a steady stream of schema migrations and feature additions through 2026.
OpenAgentd is a local-first, coding-focused AI agent platform that provides a desktop application (Tauri), mobile companion (iOS), browser-based UI (React/Vite), CLI, and an API server. The system orchestrates an agent loop with pluggable LLM providers (Anthropic, OpenAI, DeepSeek, Grok, Ollama, etc.), MCP tool integration, filesystem tools, code analysis, and a scheduler for recurring tasks. It features a dual backend architecture: a legacy Python/FastAPI implementation and a newer Rust v3 rewrite (axum-based) that is progressively replacing the Python version. The web frontend is a rich SPA with real-time streaming, workspace management, telemetry/observability dashboards, and extensive settings panels. The repository also includes code health analysis scripts, document chunking for RAG, and a comprehensive test suite across all layers.
Kubewall is a production-grade, open-source Kubernetes dashboard that provides real-time multi-cluster management through a single-binary web application. It combines a Go backend (Echo framework, client-go informers) with a React/TypeScript frontend (Redux Toolkit, shadcn/ui, TanStack Router) to offer live monitoring, YAML editing, pod log streaming, port forwarding, and an AI-powered troubleshooting assistant (kwAI). The architecture relies on Server-Sent Events (SSE) to push Kubernetes resource updates to the browser, with an in-memory informer cache per cluster. The backend exposes REST endpoints for every standard Kubernetes resource type as well as Custom Resource Definitions (CRDs), while the frontend uses virtualized data tables, Radix UI primitives, and lazy-loaded Monaco editor for editing. The system supports multiple kubeconfig authentication methods and includes a Helm chart for deployment.
The repository is a suite of tools for UniFi network management, including a dashboard, Wi-Fi Stalker, Threat Watch, and Network Pulse. It's built using FastAPI, Python 3.9+, and utilizes a multi-stage Docker build. The application has a React-based frontend and uses a modular design with separate tools for different features.
This repository implements a self-hostable data integration platform for building, deploying, and managing ETL/ELT pipelines. The system features a polyglot microservices architecture: a React frontend (app-ux) for workflow visualization and configuration, two Spring Boot backend services (mgmt-svc for management APIs and ingestion-svc for event intake), and a Node.js processor for executing pipeline transformations. Core functionality includes defining sources, transformations, and destinations; scheduling pipelines via cron; monitoring execution metrics; role-based access control with SSO support; and a plugin architecture for community-contributed apps. Data flows involve ingestion services writing events to PostgreSQL queues, the processor executing transformation steps in sandboxed VMs, and delivering results to destinations. The platform emphasizes secure handling of sensitive data via encryption utilities and provides extensibility through dynamic script execution.
This repository implements a multi-agent orchestrator system for healthcare workflows, specifically designed for cancer care/tumor board scenarios. The system coordinates specialized AI agents (Patient History, Radiology, Patient Status, Clinical Guidelines, Report Creation, Clinical Trials, Medical Research) via Semantic Kernel group chat, with an Orchestrator agent facilitating collaboration. It provides both Microsoft Teams integration and a React/Vite web interface for user interaction. The architecture includes a Python/FastAPI backend handling WebSocket connections and Bot Framework messages, agent-specific tools for data access and external API integration, and an evaluation framework for assessing agent performance. Infrastructure is provisioned via Azure Bicep templates deploying AI services, storage, compute, and networking resources.
This is a web application for viewing and analyzing Minecraft server performance timings reports. It consists of a PHP backend that handles report upload, caching, and data transformation, and a React-based frontend for interactive visualization. The backend parses a proprietary compressed timing format from Minecraft servers (Paper, Sponge, etc.) into a structured JSON object, caches it, and serves it via REST endpoints. The frontend fetches this data and renders charts, tables, and drill-down views of timing handlers, regions, and plugins. The application is a single-page application built with React 15, Webpack 2, and Babel 6, with a heavy reliance on jQuery-like PHP utilities and the Stiphle rate-limiting library. The architecture is a classic LAMP stack with PHP serving as both the API and the HTML shell. Key entry points are index.php (main page), data.php (API), post.php (upload endpoint), and backend_daemon.php (CLI daemon). The system uses a file-based cache with gzip compression and serialization. The frontend manages state with a global data object and uses Chart.js for TPS and entity graphs. The codebase demonstrates a custom JSON deserialization trait (FromJson) that maps JSON arrays to typed PHP objects using PHPDoc annotations, which is a notable pattern for handling evolving data formats. However, the project shows significant technical debt with outdated dependencies and a monolithic global state approach in the frontend.