Look, we all know GitHub has millions of repos. Finding the ones actually worth your time? That's the real challenge.
I've spent a couple hours digging through repositories, testing tools, and talking with other developers to find projects that are genuinely useful. Projects that'll keep you sharp, update your toolkit, and maybe even give you something cool to contribute to.
What makes these projects special? They're actively maintained, they solve real problems, and they welcome new contributors. Some are revolutionary AI tools that are changing how we build intelligent systems. Others are developer utilities that will make you wonder how you ever lived without them.
Whether you're looking to strengthen your resume, sharpen your skills on something new, or just give back to the community, these are the projects where your contributions will make a real impact.
Let’s get to it!
Level up your dev career! Join Index.dev, get matched with top global companies, and work on ground-breaking remote projects.
1. LangChain
LangChain is an open-source framework for building applications powered by large language models (LLMs). It's the glue that connects large language models with other sources of computation and data.
What You Can Do
- Create AI chatbots and agents that combine the reasoning capabilities of LLMs with external tools, data sources, and APIs.
- Contribute new chain types or improve integrations with LLMs.
- Automate document processing and Q&A systems.
- Build AI-powered search and recommendations.
Features
- Works with OpenAI, Hugging Face, and more
- Tools for memory management and prompt engineering
- Agents that can use tools based on their reasoning
Use Cases
LangChain powers chatbots, question-answering systems, summarization tools, and content generation applications. It's particularly valuable for applications that need to combine language models with custom data or specific domain knowledge.
2. PyTorch
PyTorch is Facebook's machine learning framework that's known for its simplicity and flexibility, especially for research. Build everything from simple neural nets to cutting-edge research models, fast. Great for both prototyping and production.
What You Can Do
- Help expand or optimize the core tensor operations (C++ or Python).
- Increase test coverage or fix flaky tests to keep everything stable.
- Profile, tune, or rewrite slow components for faster training and inference.
- Work on autograd, distributed training, quantization, or mobile support.
- Develop new modules, layers, or integrations with external libraries.
Features
- Dynamic computational graph (define-by-run)
- Native CUDA support for GPU acceleration
- TorchScript for production deployment
- Strong Ecosystem: Hugging Face, fastai, and thousands more use it
- First-class support for computer vision and NLP tasks
Use Cases
PyTorch excels in research settings and is the framework of choice for many academic papers. It's great for computer vision, NLP, and reinforcement learning projects.
3. Open R1
Open R1 is Hugging Face's ambitious project to reproduce and extend the DeepSeek-R1 pipeline in an open-source environment. It's democratizing cutting-edge AI research by making advanced reasoning, math, and coding models accessible to everyone.
What You Can Do
- Contribute to training models, evaluation scripts, or synthetic data generation.
- Strengthen the RL pipeline or optimize multi-stage training processes.
Features
- Scripts for supervised fine-tuning (SFT), reinforcement learning (RL), and synthetic data generation
- Modular design for easy experimentation and collaboration
- Streamlined execution through simple Makefile commands
Use Cases
Open R1 empowers researchers and developers to explore advanced AI topics like RL and synthetic data, making AI accessible to everyone. It's being used for exploring reinforcement learning techniques, creating synthetic training data, and building more capable AI systems that excel at reasoning tasks.
4. Turborepo
Turborepo is a high-performance build system for JavaScript and TypeScript monorepos, built by the folks at Vercel. It helps you manage multiple packages/apps in a monorepo setup without slow build times or tangled scripts.
What You Can Do
- Optimize caching and task scheduling logic.
- Improve CI/CD integration.
- Write and refine tests ensuring stability with better bug fixes.
- Add support for more languages.
Features
- Incremental builds with intelligent caching
- Parallel task execution
- Remote caching for sharing build artifacts across machines
- Zero configuration required for many common setups
- Seamless integration with npm, yarn, and pnpm workspaces
Use Cases
Turborepo has become essential for teams managing complex JavaScript projects, especially those with multiple packages or applications sharing common code. It's being used by companies with large frontend codebases to cut build times.
5. Zig
Zig is a programming language focused on simplicity and performance. It’s designed as an alternative to C/C++ with modern features like safety and cross-compilation.
What You Can Do
- Write high-performance and low-level code.
- Improve compiler optimizations or add standard library features.
- Contribute to language tooling like debuggers or formatters.
- Use it as a drop-in replacement for C.
Features
- Manual memory management with safety checks. Like Rust, but simpler
- Cross-compilation built into the compiler
- No hidden control flow and memory allocations
- Built-in testing framework
Use Cases
Zig is gaining traction for system-level programming, game development, and embedded systems development where performance is critical. It's ideal for developers who need C-like performance but want to avoid common pitfalls and security vulnerabilities.
6. Supabase
Supabase is an open-source alternative to Firebase. It provides backend services you need to build a product like authentication, real-time databases, and storage out of the box.
What You Can Do
- Build full-stack apps with a real-time Postgres database.
- Handle authentication, storage, and APIs easily.
- Get Firebase-like developer experience but open-source.
- Contribute features like advanced analytics or caching layers.
- Add new SDKs or improve existing ones.
Features
- PostgreSQL database with real-time capabilities
- Easy integration with front-end frameworks
- Authentication with multiple providers
- Auto-generated APIs (REST and GraphQL)
- Edge Functions for serverless computing
Use Cases
Supabase is ideal for startups and independent developers building web and mobile applications. It's being used for social platforms, SaaS products, e-commerce applications, and any project that needs a scalable backend without vendor lock-in.
7. PySyft
PySyft is an open-source Python library that lets you build privacy-preserving machine learning models. It’s built by the OpenMined community and helps you train models on data without ever seeing the data.
What You Can Do
- Develop integrations with deep learning frameworks like PyTorch and TensorFlow.
- Work on encryption techniques for secure AI training.
- Spot inefficiencies and write unit tests to keep the core stable and future-proof.
Features
- Works with PyTorch and JAX
- Secure multi-party computation
- Federated learning infrastructure
- Encrypted machine learning operations
Use Cases
PySyft allows AI models to be trained on sensitive data (like healthcare records) without exposing the actual data, making AI safer for industries like finance and healthcare.
8. Base Web
Base Web is a highly customizable React UI framework built by Uber for designing enterprise-scale websites and web apps. It helps developers build responsive, accessible, and fast UIs with ease.
What You Can Do
- Design and build new UI components.
- Improve styling and theme capabilities.
- Optimize performance for large-scale apps.
- Write and run tests to ensure component functionality.
Features
- 50+ production-ready components
- Fully themeable with design tokens
- Works seamlessly with React
- First-class TypeScript support
- Server-side rendering compatible
Use Cases
Base Web shines in enterprise environments where reliability and consistency are crucial. It's ideal for large-scale web applications, design systems, and any project where you need professional components without starting from scratch.
9. Deno
Deno is a secure runtime for JavaScript and TypeScript created by the original developer of Node.js, focusing on security and developer experience. It's built to fix the mistakes and design flaws he saw in Node. With Deno, you can write server-side JavaScript that's secure by default, with no need for package.json or node_modules hell. Just write code and run it.
What You Can Do
- Contribute to build tools, IDE integrations, or third-party module ecosystem.
- Help expand or improve the Deno standard library with new modules.
- Improve the docs to make Deno easier to learn and use for new developers.
- Identify and resolve issues in the Deno runtime or ecosystem.
Features
- Security first: No file, network, or environment access without explicit permission
- Built-in TypeScript support without configuration
- Standard library of reviewed, high-quality code
- Single executable with no external dependencies
- Compatible with browser APIs
Use Cases
Deno is perfect for building secure web applications, command-line tools, and server-side rendering. It's especially valuable for projects where security is a priority.
10. MIND (by Index.dev)
MIND Tech AI is a massive, AI-ready, open-source knowledge base that solves one of tech recruiting’s biggest headaches: matching skills accurately across different terminologies and frameworks.
What You Can Do
- Improve job-to-candidate matching in hiring platforms.
- Train ML models to “get” tech skills contextually.
- Power better job recommendation engines.
Features
- Comprehensive mapping of tech skills, frameworks, and concepts
- Built to help LLMs and recommender systems understand tech concepts
- Implied skill connections (knowing Next.js implies React knowledge)
- AI-ready data structure for integration into matching algorithms
Use Cases
MIND is a game-changer for recruitment platforms, HR tools, and skill-matching applications. It's perfect for developers building career sites, talent marketplaces, or any application that needs to understand tech skills beyond simple keyword matching. If you've ever been frustrated by the differences between "ML" and "Machine Learning" in search results, this project is solving that problem at scale.
Improve skill matching, add new tech skills, and help train AI. Fork and contribute to MIND’s tech skills ontology on GitHub!
11. Airbyte
Airbyte is an open-source data integration platform that syncs data from applications, APIs, and databases to warehouses and lakes.
What You Can Do
- Build new connectors to expand Airbyte's integration capabilities.
- Strengthen performance, fix bugs, or add features to current connectors.
- Update or create documentation to assist users and contributors.
- Identify issues and submit pull requests with fixes.
- Improve the user interface and experience of the Airbyte platform.
Features
- 300+ pre-built connectors
- Plug-and-play dbt transformations
- Custom connector SDK
- Built-in orchestration and scheduling
- Data quality monitoring
Use Cases
Airbyte is essential for data teams that need to consolidate information from multiple sources. It's being used for business intelligence, customer data platforms, marketing analytics, and any scenario where data needs to flow between systems reliably.
12. Rath
Rath is an open-source library designed for building data pipelines in Python. It simplifies the process of connecting APIs, transforming data, and managing workflows.
What You Can Do
- Contribute to Rath’s core features, such as user management, workflows, and more.
- Add or improve integrations with other tools and services like APIs, databases, or messaging tools.
- Help identify and fix bugs in the codebase (automation logic, UI glitches, and performance bottlenecks).
- Write tests to ensure Rath works across different environments and setups.
Features
- Automated data analysis and insight discovery
- Smart visualization recommendations
- Interactive dashboard creation
- Data cleaning and transformation tools
- Support for various data formats (CSV, JSON, Excel)
- No-code interface with advanced customization options
Use Cases
Great for ETL processes, integrating third-party APIs, or automating data workflows in analytics projects.
13. Payload CMS
Payload is a headless CMS and full app framework rolled into one. Built on TypeScript, React, Express, and MongoDB it gives you total control over content modeling, access control, and API structure. It’s self-hosted, fully extensible, and actually enjoyable to work with.
What You Can Do
- Build new features like better API support or developer tooling.
- Fix bugs in the backend or API layers.
- Create custom field types or plugins.
- Help keep the documentation clear and up-to-date.
- Optimize performance through better queries and smarter caching.
Features
- Self-hosted and completely customizable
- GraphQL and REST APIs out of the box
- Authentication and access control
- Admin panel built with React
- Built-in versioning & localization
Use Cases
Perfect for e-commerce, SaaS dashboards, marketing platforms, or any app that needs flexible, structured content management with powerful APIs.
14. RLHF + PaLM
This project focuses on Reinforcement Learning with Human Feedback (RLHF) applied to Google’s PaLM (Pathways Language Model). It aims to align large language models with human values and preferences.
What You Can Do
- Fine-tune AI models to align with human preferences.
- Reduce bias and hallucinations in AI responses.
- Build better AI assistants and chatbots.
- Contribute datasets or improve training pipelines.
Features
- Human-in-the-loop training for ethical AI development
- Advanced natural language understanding capabilities
- Used by Google’s PaLM
Use Cases
AI models trained with RLHF are crucial for creating less biased and safer AI systems that align with user needs in applications like virtual assistants or content moderation.
15. Tauri
Tauri is a toolkit for building lightweight, secure desktop applications with web technologies (like HTML, CSS, JS) + Rust. You can create desktop apps for all major platforms with smaller binaries than Electron.
What You Can Do
- Help improve the Rust backend or refine the cross-platform bindings.
- Find and fix issues across Windows, macOS, and Linux.
- Build and maintain plugins to extend Tauri’s native functionality.
- Test apps across environments and improve performance.
Features
- Tiny application bundle size (often less than 10MB)
- Use any front-end framework (React, Vue, Svelte, etc.)
- Build for Windows, macOS, and Linux
- Robust security defaults
- Access to system APIs through Rust
Use Cases
Tauri is perfect for developers who want to create desktop applications with web technologies but are concerned about the resource usage and security models of other frameworks. It's great for tools, utilities, and productivity applications.
16. Qdrant
Qdrant is an open-source vector search engine optimized for neural search applications. It’s designed for high-performance similarity searches—like images, text, or audio—in machine learning systems.
What You Can Do
- Improve indexing algorithms or add integrations with ML frameworks.
- Create AI-powered recommendation systems.
- Contribute to API improvements or scalability features.
- Store and query large-scale embeddings for machine learning models.
Features
- High-speed vector similarity search
- Scalable architecture with multi-node clustering
- REST API and gRPC interfaces
- Optimized for production use with monitoring and metrics
Use Cases
Qdrant powers semantic search, recommendation systems, and AI applications that need to find "similar" items efficiently. It's ideal for e-commerce product discovery, content recommendation systems, and image search applications.
17. NocoDB
NocoDB is like Airtable, but open-source. It takes any database (MySQL, PostgreSQL, etc.) and gives it a beautiful, spreadsheet-like UI. It’s a no-code platform that helps developers manage databases visually while offering customization through APIs.
What You Can Do
- Create powerful no-coding applications with robust databases underneath.
- Add new integrations or improve plugin architecture.
- Contribute to UI/UX improvements or performance optimization.
Features
- Spreadsheet-style interface for databases
- API-first design for automation and integrations
- Automations and webhooks
- Integration with third-party apps
Use Cases
Perfect for teams managing CRM, inventory systems, or project tracking without writing extensive code. It's also being used as a backend for simple web applications.
18. HeroUI
HeroUI (previously NextUI) is an open-source UI component library tailored for developers who want modern design and clean code. You can create stunning web applications with accessible, customizable components that work perfectly with React and Next.js.
What You Can Do
- Design sleek, accessible, and responsive React components.
- Polish the UI/UX and help squash layout or style issues.
- Write tests to help ensure components behave as expected.
- Help make NextUI work better across languages and regions.
Features
- Fully accessible components
- Fully themeable with modern UI patterns
- Optimized for server-side rendering (SSR)
- Lightweight and responsive
Use Cases
HeroUI makes it easy for developers to build polished UIs without bloated styles or complex setups, improving user experience across modern web apps. It's being used for dashboards, consumer-facing websites, and admin interfaces where both aesthetics and performance matter.
19. Dokku
Dokku is a self-hosted PaaS (Platform as a Service) that lets you deploy and manage apps just like Heroku, but on your own servers.
What You Can Do
- Add plugins for additional functionality.
- Improve deployment workflows or documentation.
- Deploy apps with a simple git push command.
- Run multiple services like databases, workers, and web apps.
Features
- Git-based deployments
- Plugin ecosystem for extensibility
- Runs on any cloud or server
Use Cases
Ideal for developers who want Heroku-like simplicity but prefer self-hosted solutions to save costs.
20. Gogs
Gogs is like GitHub, but you host it yourself. It’s a super lightweight, self-hosted Git service written in Go that runs almost anywhere.
What You Can Do
- Add new features or improve scalability.
- Fix bugs or improve documentation.
- Collaborate with your team without relying on GitHub or GitLab.
- Keep full control over your repositories and data.
Features
- Minimal system requirements
- Cross-platform support
- Works on Windows, Linux, macOS
- Runs even on a Raspberry Pi
Use Cases
Gogs is ideal for small teams needing a simple, self-hosted Git server without heavy infrastructure requirements.
Explore More: Top 18 Open-Source Apps Every Developer Should Know in 2025
Ready to Make Your Mark?
These 20 projects aren't just repositories on GitHub; they're communities waiting for you to jump in.
Whether you're into AI with Open R1, powering better job recommendation engines with MIND, or creating no-code tools with NocoDB, there's something here that matches your skills.
And the best part? Contributing to these projects it’s about building your reputation, learning from brilliant developers worldwide, and solving problems that actually matter.
Pick a project that excites you. Start small – fix a bug, improve documentation, or add a feature. Before you know it, your GitHub profile will showcase meaningful contributions that employers actually care about.
Remember: today's side project could be tomorrow's industry standard. And you could be the one who helped build it.
So what are you waiting for? Fork, clone, and make your first pull request.
Join Index.dev's talent network today and get matched with companies that value your open source contributions—because developers who build the future deserve the best opportunities.
