CockroachDB is a cloud-native distributed SQL database built on a transactional, strongly-consistent key-value store (Pebble). It provides horizontal scalability, survivability against disk/machine/rack/datacenter failures with minimal latency disruption, strongly-consistent ACID transactions, and a PostgreSQL-compatible SQL API. The architecture is a monolithic Go application with a clear separation between the SQL layer (parser, optimizer, distributed execution engine) and the KV layer (Raft-based replication, Pebble storage). Key innovations include a Cascades-style cost-based optimizer with memoization, a vectorized columnar execution engine (colexec), online declarative schema changes, distributed backup/restore with incremental support, change data capture with multiple sink formats, and multi-region/multi-tenant capabilities with geo-partitioning. The codebase demonstrates exceptional engineering craftsmanship throughout - the Raft implementation, distributed SQL planner, admission control system, and time-series storage are all genuinely reusable components that showcase non-obvious solutions to hard distributed systems problems. Active development with recent commits, comprehensive CI/CD, and regular releases.
Metabase is an open-source business intelligence platform that enables users to query, visualize, and share data through an intuitive interface. The system is built as a full-stack application with a Clojure backend and React/TypeScript frontend, supporting both self-hosted and cloud deployments. The architecture follows a modular monolith pattern with clear separation between the core application, enterprise features, and embedding SDK components. The codebase demonstrates sophisticated engineering practices including a comprehensive plugin system that allows enterprise features to extend core functionality, a custom query builder with both GUI and SQL interfaces, and a robust embedding SDK for integrating Metabase components into external applications. The frontend utilizes modern React patterns with Redux for state management, while the backend employs Clojure's functional programming strengths for data processing and API handling. Key architectural highlights include the metabase-lib library for query construction and metadata management, a sophisticated visualization system with ECharts integration, and a multi-tenant embedding solution with JWT/SAML authentication. The codebase shows strong attention to internationalization, accessibility, and performance optimization through features like query caching, lazy loading, and bundle splitting.
ERPNext is a comprehensive open-source ERP (Enterprise Resource Planning) system built on the Frappe Framework. It provides a full suite of business management modules including Accounting, CRM, Buying, Selling, Stock/Inventory, Manufacturing, Projects, Assets, Quality Management, Support, and more. The system uses a metadata-driven architecture where DocTypes (document types) define data models, business logic, and UI behavior. The codebase follows a modular structure with each business domain in its own directory (accounts, buying, selling, stock, manufacturing, projects, etc.). It includes a modern React/TypeScript banking module built with Vite for bank reconciliation and statement importing. The system supports multi-company, multi-currency, multi-language deployments with extensive regional/localization support for tax compliance across many countries. It has a robust patch/migration system for database schema evolution and a comprehensive test suite.
Appwrite is an open-source backend-as-a-service (BaaS) platform that provides a comprehensive set of APIs for authentication, databases, storage, serverless functions, messaging, realtime, and more. The codebase is a hybrid monolithic-microservice architecture: a central PHP application (using the Swoole coroutine-based HTTP server) handles REST, GraphQL, and WebSocket requests, while background workers (over 20 types) process asynchronous tasks such as email delivery, function execution, database migrations, and usage analytics. The platform supports multiple database backends (MariaDB, MongoDB, PostgreSQL, ClickHouse) and a variety of messaging providers (SMS, email, push). It is designed for multi-tenant deployments with per-project isolation, OAuth2 social login, and extensive API versioning. The repository also contains a large collection of reusable PHP packages (the Utopia framework) that provide caching, queueing, query building, DNS resolution, NATS messaging, HTTP client, PSR-7 implementation, and more. These packages are independently published and can be extracted for use in other projects.
LibreChat is a self-hosted, multi-tenant AI chat platform that unifies access to major LLM providers (OpenAI, Anthropic, Google, AWS Bedrock, Azure, and custom OpenAI-compatible endpoints) through a single web interface. The architecture is a monolithic Express.js/Node.js backend serving a React (Vite) SPA, backed by MongoDB (Mongoose), Redis for caching/rate-limiting, MeiliSearch for full-text search, and PostgreSQL (pgvector) for RAG embeddings. The codebase is organized as a monorepo with shared packages: `data-schemas` (Mongoose models/methods with multi-tenant isolation), `data-provider` (Zod schemas, API types, React Query hooks), `api` (MCP management, streaming, auth, SSRF protection), and `client` (reusable React component library). Key data flows include: user requests hitting Express routes, middleware performing JWT/OAuth/LDAP/SAML auth and RBAC checks, agent orchestration via `@librechat/agents` with tool execution (MCP, code interpreter, web search), and SSE streaming with resumable stream support via Redis-backed job management. The system implements a sophisticated multi-tenant isolation layer using Mongoose plugins (`tenantIsolation`) that automatically scope queries by tenant, an ACL/permissions system with bit-flag-based permission checks, and SSRF protection with DNS resolution validation for MCP server URLs. The MCP (Model Context Protocol) subsystem is particularly well-developed, featuring OAuth flow management with reconnection tracking, tool discovery, and both stdio and HTTP/SSE transport support. The streaming architecture supports horizontal scaling with Redis-backed event transport and job stores, enabling resumable streams across multiple server instances.
Monica is an open source personal relationship management system (PRM) built with Laravel and Vue.js that functions as a 'CRM for your friends and family'. The system features comprehensive contact management, relationship tracking, activity logging (notes, calls, tasks), life event tracking, document management, and journaling capabilities. It implements a modular domain-driven architecture with clear separation between backend services, view helpers for data preparation, and frontend components. The application includes CardDAV synchronization for external contact sync, multi-vault/user support, notification systems (email/Telegram), reporting features, and extensive internationalization (27+ languages). Built with modern Laravel practices including service contracts, queueable jobs, event-driven architecture, and comprehensive testing.
aiproxy is a production-grade AI API gateway written in Go, acting as a unified proxy between client applications and over 40 AI provider APIs (OpenAI, Anthropic, Gemini, etc.). It converts protocol formats (OpenAI, Anthropic, Gemini), handles authentication, rate limiting, load balancing, request caching, and detailed usage billing. The system includes a React admin dashboard for managing channels, tokens, models, groups, logs, and consumption. It also supports the Model Context Protocol (MCP) for integrating AI tool servers, both hosted and proxied. The architecture is monolithic with a plugin system for extensibility, storing state in PostgreSQL and Redis. It demonstrates strong engineering practices including comprehensive test coverage, efficient caching (multi-layer with keyed locks), sliding window rate limiting, and a clean adaptor pattern for provider abstraction.
Tandoor is a self-hosted, multi-tenant recipe management web application. The system is a monolithic Django application (cookbook app) exposing a Django REST Framework API, paired with a Vue 3 + Vuetify single-page frontend that consumes an auto-generated TypeScript OpenAPI client. Core production domains, verified in code, are: structured recipes (steps, ingredients, scaling), a tree-based hierarchical food database with nutritional properties and USDA FDC IDs, meal planning with iCal export, shopping lists with supermarket categories and offline sync queues, inventory/pantry booking, cookbooks, share links, and view/cook logging. Tenancy is enforced through Space/Household/UserSpace models with django-scopes middleware and granular guest/user/admin permission classes rather than through database row-level security. The largest subsystem is recipe import/export: an abstract Integration base class with 20+ adapters for other recipe managers (Paprika, Mealie, Cooklang, Gourmet, ChefTap, etc.), URL scraping via recipe-scrapers/microdata/pytubefix, cloud storage sync providers (Dropbox, Nextcloud, WebDAV, local), a bulk open-data importer for USDA-derived reference foods/units/conversions, and an embedded PDF.js viewer for external recipe files. A connector/event layer dispatches async updates to external services such as Home Assistant, a Telegram bot webhook writes shopping list entries, and recent AI features (litellm-based, per-space credit tracked) handle image/PDF recipe import, step sorting, and nutrition extraction. Background work is done via threads rather than a task queue. Craftsmanship is above average: clean abstract patterns (Integration, Provider, Connector ABCs), chainable queryset managers, mixin-based permission/merge/tree logic, a strong permission-matrix test suite, and genuinely reusable helper algorithms (BFS unit conversion, ingredient parsing, Cooklang parsing). However, the codebase carries notable debt: 240+ migrations with heavy schema churn, a ~3000-line monolithic api.py view file, committed auto-generated OpenAPI client code that can drift, vendored PDF.js binaries, some tests that suppress tenant scoping via scopes_disabled, and copy-paste duplication inside the open-data importer. Overall it is a well-maintained, actively developed application with solid architecture for its scale.
OrchardCore is a mature, open-source modular and multi-tenant application framework and CMS built on ASP.NET Core. The codebase reveals a production-grade system with a rich extensibility model: content parts and fields with display drivers, workflow engine with activities, deployment/recipe system, multi-tenant shell contexts, and multiple search backends (Lucene, Elasticsearch, Azure AI). The frontend admin UI is implemented in Vue 3 with TypeScript (the 'Bloom' framework), providing reusable UI components and a data pipeline designer. Under the hood, it uses YesSql as a document ORM, SignalR for real-time notifications, and supports various authentication providers, localization, and media management. Despite its large surface, the architecture is cleanly separated into modules with consistent patterns (display drivers, handlers, permission providers), though some legacy deprecations and duplicate search provider code indicate manageable technical debt.
Apache Polaris is a production-grade, multi-tenant Iceberg REST catalog server built on Quarkus. It provides a REST API for catalog management (namespaces, tables, views), credential vending for cloud storage (S3, GCS, Azure), fine-grained authorization via built-in RBAC or external OPA/Ranger, and event publishing (Kafka, CloudWatch). The system supports two persistence backends: a legacy relational JDBC (PostgreSQL, H2) and a new NoSQL abstraction with custom indexing, caching, and distributed invalidation. Multi-tenancy is implemented via isolated realms, each with its own configuration and persisted state. Key entry points are the Quarkus server (`polaris-server`), a CLI admin tool, and a Python client. External dependencies include Iceberg libraries, Quarkus, Jackson, Testcontainers, and optional integrations like Open Policy Agent, Apache Ranger, and Kafka. The architecture is modular with clear SPI boundaries, making extensions (persistence, auth, federation) pluggable.
FastGPT is a production-grade AI Agent building platform that enables users to create, deploy, and manage AI applications through a visual workflow editor. The system provides a full-stack solution including a Next.js frontend with Chakra UI, a comprehensive REST API layer, and a modular backend architecture. Core capabilities include a workflow engine with agent loops, tool calling, dataset (knowledge base) management with vector search, sandboxed code execution (JS/Python), multi-tenant team management with fine-grained permissions, billing and usage tracking, and integration with external platforms (WeChat, Feishu, Dingtalk, MCP). The platform is designed for both open-source community and commercial deployments, with feature gating via license activation. The codebase is organized as a monorepo with clear separation of concerns: shared types and constants in `packages/global`, business logic in `packages/service`, UI components in `packages/web`, data access layer in `packages/dal`, and separate services for sandbox execution and volume management. It relies on MongoDB, Redis, BullMQ, S3-compatible storage, and multiple vector database backends (PG, Milvus, OceanBase, SeekDB, OpenGauss). The architecture is predominantly monolithic with the main application as a Next.js app, but sandbox and volume manager run as independent services, indicating a hybrid monolith/microservices pattern.
Bagisto is a production-ready open-source eCommerce platform built on Laravel with a modular monolith architecture. It provides a complete admin panel and storefront for managing multi-channel, multi-locale, multi-currency online stores. The system supports a wide range of product types (simple, configurable, bundle, grouped, downloadable, virtual, booking), payment gateways (Stripe, PayPal, Razorpay, PayU, PhonePe, PayGlocal), marketing promotions (cart rules, catalog rules), full-page caching, Elasticsearch integration, GDPR compliance, EU withdrawal management, RMA (returns), data import/export, and AI-powered features (MagicAI). The architecture is event-driven with listeners for cache invalidation, notifications, and indexing. The admin panel uses Vue 3 components with Blade templates, while the storefront is built with Tailwind CSS and Vue. The codebase is organized into ~40+ packages under packages/Webkul/, each with its own models, repositories, migrations, translations, and tests. Key entry points include the admin routes (packages/Webkul/Admin/src/Routes/web.php), shop routes (packages/Webkul/Shop/src/Routes/), and the installer wizard. The system relies on MySQL, Redis, Elasticsearch, and various external payment/shipping APIs. The core framework (Concord) provides facades for core services, ACL, menu, system config, and Elasticsearch. The DataGrid system offers a reusable abstraction for listing and filtering entities across the admin panel. Notable engineering highlights include the CurrencyFormatter trait for locale-aware currency display, the MediaFileName helper for safe file upload handling, the ForgetsPages trait for full-page cache invalidation, and the DynamicMailTransport for configurable email delivery. The event-driven architecture with listeners and the repository pattern are consistently applied across all modules.
JeecgBoot is an enterprise-grade low-code development platform that accelerates application delivery through AI-powered code generation, online form/report builders, and workflow integration. The backend is built on Spring Boot 3 with MyBatis-Plus, Shiro, and JWT, supporting both monolithic and microservice architectures (Spring Cloud, Nacos, Sentinel, XXL-Job). The frontend uses Vue 3 + TypeScript + Ant Design Vue with a rich component library including JVxeTable for editable grids and schema-driven forms. The platform provides comprehensive RBAC, multi-tenancy, dynamic data sources, data dictionaries, and third-party integrations (DingTalk, WeChat, Feishu). A dedicated AI module (airag) offers RAG knowledge bases, LLM chat, voice/image generation, and MCP plugin support. The codebase is actively maintained with frequent commits and a large community, though it contains some commented-out legacy code and hardcoded security values that represent minor technical debt.
Twenty is a production-grade, open-source CRM delivered as a large Nx monorepo. The system is a full-stack TypeScript application: a NestJS backend (twenty-server) exposing GraphQL/REST APIs, a React frontend (twenty-front) with Jotai-based state management, a shared UI component library (twenty-ui), and shared TypeScript packages (twenty-shared). It provides multi-tenant workspaces, object-record CRUD with custom data models, role-based permissions, AI chat/agent capabilities (model tiers, tool calls, streaming), analytics via ClickHouse, and a CLI scaffolding tool (create-twenty-app) for bootstrapping custom apps. The architecture is modular with NestJS ConfigurableModuleBuilder patterns, event-driven design using Zod-validated schemas, and component-scoped state via atom families and component instance contexts.
ZaloCRM is a production-grade, multi-tenant customer relationship management system deeply integrated with the Vietnamese Zalo messaging platform. It serves real-estate sales teams by unifying contact management, real-time chat, lead scoring/automation, appointment scheduling, media library, RBAC, privacy controls, and third-party integrations (Facebook Lead Ads, Telegram bridge, Google Sheets). The system is deployed as a Node.js/Fastify monolith with PostgreSQL (Prisma ORM), Redis (BullMQ queues, rate limiting), MinIO/S3-compatible storage, and Socket.IO for real-time events. It relies on the unofficial Zalo SDK (zca-js) for messaging operations and supports multiple AI providers (Anthropic, Gemini, OpenAI) for reply drafting, sentiment analysis, and appointment parsing. The architecture follows an open-core model: the Community edition ships with extension stubs (`_ee-stubs`) that are replaced in the Enterprise edition. Key entry points include auth routes (JWT with rotating refresh tokens), Zalo account management (QR login, reconnect), friend sync (cron-driven diff-then-emit), and a rich chat operations module (reactions, edits, forwards). Background jobs handle lead pool assignment, care session monitoring, scoring decay, and engagement pattern classification. The system maintains strong data invariants through Prisma extensions (tenant context, RLS preparation, safe-contact-write) and a custom privacy layer that redacts private-nick messages based on viewer role. Data flows center on the Zalo account pool: inbound messages are ingested via socket listeners or friend sync, routed to conversations via a global-aware resolver, and triggers automation hooks (care sessions, auto-tags, sequence advancement). Analytics are derived from daily engagement materialization and scoring engine (4 dimensions with weights). External dependencies are managed via Redis-backed rate limiters, while storage operations are abstracted behind a `StorageDriver` interface with local/R2 implementations. The codebase is actively maintained with daily commits, extensive Prisma migrations, and a comprehensive test suite covering privacy invariants, merge policies, and regression scenarios.
ruoyi-vue-pro (芋道) is a large-scale, multi-module enterprise rapid development platform built on Spring Boot 2.7, MyBatis-Plus, and Vue 3. It provides a comprehensive backend infrastructure for building business applications such as ERP, CRM, HRM, MES, WMS, OA, BPM (Flowable), IoT, AI, IM, Pay, Mall, and more. The codebase is structured around a reusable framework layer (`yudao-framework`) with many Spring Boot starters for security, data permission, multi-tenancy, WebSocket, Excel, monitoring, and protection (rate limiting, idempotency). Business modules follow a consistent layered pattern (DO → Mapper → Service → Controller → VO) with MapStruct converters. The system supports MySQL, PostgreSQL, Oracle, SQL Server, and Dameng databases, and integrates with Redis, RabbitMQ, Kafka, RocketMQ for messaging. It includes a code generator, a workflow designer, and AI chat/image/mind-map capabilities. The architecture is modular and well-organized, though the sheer number of modules (20+) introduces some duplication and complexity. Overall, it is a mature, actively maintained project with strong adherence to common enterprise patterns.
WxJava is a comprehensive Java SDK for the WeChat ecosystem, providing client libraries for Official Accounts (MP), Mini Programs (MiniApp), Enterprise WeChat (CP), WeChat Pay, Open Platform, Video Channels, and more. It is designed as a multi-module library to be embedded in Java applications, with first-class support for Spring Boot and Solon frameworks via auto-configuration starters. The SDK abstracts HTTP client backends (Apache, OkHttp, Jodd, HttpComponents), provides Redis-based config storage for distributed token management, and implements multi-tenant patterns for managing multiple WeChat accounts. It handles the full lifecycle of WeChat API integration: request signing, serialization (Gson/Jackson/XStream), message routing, and callback verification. The codebase is large and well-structured, though it shows signs of duplication across modules and mixed serialization frameworks, reflecting organic growth over time.
Onyx is a full-stack AI enterprise search and chat platform. The backend is a Python/FastAPI service with Celery workers, PostgreSQL, Redis, and a vector search engine (Vespa/OpenSearch). It ingests documents from 50+ external connectors (Google Drive, Slack, Confluence, Jira, SharePoint, etc.), runs a chunking/embedding/indexing pipeline, and serves RAG chat and search to multiple frontends: Next.js web app, React Native/Expo mobile app, Tauri desktop wrapper, Go CLI, an embeddable web widget, and a Terraform provider. It also includes an MCP server, deep research flows, sandboxed code-execution sessions (Craft), prompt caching, SSO/OAuth/SCIM multi-tenancy, rate limiting, and an admin analytics console. The repository has a hybrid character because it contains substantial data-engineering machinery (indexing pipelines, chunkers, embedding providers, evals) but the dominant identity is a production application rather than a research project.
This repository is a full-stack, multi-tenant AI API gateway platform (named 'sub2api') that proxies requests to a wide range of AI model providers including OpenAI, Anthropic, Gemini, Grok, DeepSeek, and many others. It provides authentication (JWT, OAuth, API keys), rate limiting, token-level billing, subscription management, user and admin management, payment processing (Alipay, WeChat Pay, Stripe, Airwallex, EasyPay), content moderation, channel monitoring, and operational dashboards. The backend is written in Go using the Gin framework and Ent ORM, with PostgreSQL and Redis for data storage and caching. The frontend is built with Vue 3, TypeScript, and Pinia stores, offering both admin and user interfaces. The system is designed for production deployment with Docker, includes extensive unit and integration tests, and supports a plugin architecture for extensible OAuth transport. Despite its large codebase and many features, the architecture follows a clear service-repository pattern with well-isolated packages, making it a solid example of a complex API gateway implementation.
Onyx is a comprehensive enterprise AI platform that provides knowledge retrieval, chat, and document indexing across 50+ third-party sources. The codebase reveals a FastAPI backend with Celery workers, PostgreSQL with Alembic migrations, OpenSearch vector search, Redis caching, and a separate model server. The frontend is a Next.js application with a custom design system (opal), alongside React Native mobile app, Go CLI, Chrome extension, and an embeddable widget. The platform supports deep research, custom agents, MCP servers, sandboxed code execution (Craft), and enterprise features like SSO, SCIM, billing, and multi-tenancy. Architecture is well-structured with clean separation of concerns, but the large codebase shows some drift: platform duplication across web/mobile/CLI, extensive Alembic migration history, and lint debt. The connector framework and design system are high-quality reusable components.