Welcome to The Weekly Five - your curated list of 5
exceptional open source projects I discovered this week.
The best tools are the ones you can see through. This week, we spotlight five open-source projects that hand control back to developers and users by replacing opaque services with transparent, self-hosted alternatives. From privacy-first analytics to AI-powered computer automation, these repos prove you do not have to sacrifice capability for clarity.
Here is what stands out across all five projects.
Top takeaways
Privacy and transparency are becoming table stakes: cookie-free analytics and self-hosted integrations are no longer edge cases, and projects like umami (38k+ stars) and nango show growing production adoption.
Cross-platform consistency matters more than ever: whether syncing ebooks or spinning up test containers, these tools prioritize seamless experiences across operating systems.
AI agents are moving from black-box APIs to open infrastructure: the new generation of computer-use tools lets you train, benchmark, and deploy on your own terms.
Who this issue is for
Developers and technical teams who want to replace vendor lock-in with self-hosted, auditable alternatives they can customize and control.
1. umami
Why this made the cut: A mature, privacy-first analytics platform that proves you can get Google Analytics-level insights without cookies or surveillance.
Why it matters
Traditional analytics tools collect more data than most teams need (or want to justify under GDPR). Umami flips the model: lightweight tracking, no cookies, and full ownership of your data. Self-host it or use their cloud, but either way, the code is open for inspection.
Key features
Cookie-free tracking: Compliant by default, no consent banners required
Unified dashboard: Traffic, campaigns, conversions, and revenue in one view
Audience segmentation and cohort analysis: Slice your data without exporting to a third-party tool
User journey mapping: Understand how visitors move through your site
Lightweight embed: Minimal performance impact on your site
How to use
Clone the repo and configure your database connection (PostgreSQL or MySQL supported)
Run the Docker Compose setup or deploy to your preferred Node.js host
Add the single-line tracking script to your site
Access the dashboard to view real-time analytics and build custom reports
🔗 View on GitHub | GitHub stars: 38,249
Learning resources
Self Hosting Umami Analytics Full Tutorial - Snehasis Ghosh - Walkthrough covering self-hosted Umami setup, database configuration, and dashboard customization.
2. readest
Why this made the cut: A polished, cross-platform ebook reader that syncs your library and reading progress without handing your data to a Big Tech provider.
Why it matters
Most ebook readers lock you into a single ecosystem. Readest supports EPUB, MOBI, PDF, and more, with sync options via WebDAV or OPDS so you own your library. The Tauri-based architecture keeps the app fast and native-feeling on every platform.
Key features
Format support: EPUB, MOBI, PDF, and Calibre library integration
Cross-platform sync: Use WebDAV, OPDS, or the KOReader plugin to keep progress in sync across all supported devices
Text-to-speech (TTS): Built-in audio narration for accessibility
Customizable reading experience: Themes, fonts, and layout controls
How to use
Download the latest release for your platform from the GitHub releases page
Import your ebook library or connect to your Calibre server
For cross-device sync, set up WebDAV in the app settings or install the KOReader plugin, this lets you start reading on your Mac and pick up exactly where you left off on your iPhone or iPad
Customize reading settings (fonts, themes, layout) to your preference
Use TTS for hands-free reading during commutes or chores
🔗 View on GitHub | GitHub stars: 23,544
Learning resources
Sync annotations & highlights between Readest & KOReader - Stefan Svartling - Tutorial showing how to sync reading annotations and highlights between the Readest app and KOReader e-readers.
3. nango
Why this made the cut: A unified integration platform that lets you connect to 900+ APIs with AI assistance, without building auth flows or token management from scratch.
Why it matters
Building integrations means wrestling with OAuth, token refresh, and API quirks for every service. Nango abstracts the painful parts: you write integration logic in code, and the platform handles authentication, token storage, and retries. The result is fewer security footguns and faster time to production.
Key features
OAuth1 and OAuth2 support: Pre-built flows for most major APIs
Token management: Automatic refresh and secure storage
AI-assisted integration building: Describe the integration you need in natural language and Nango generates TypeScript sync scripts, maps API fields, and suggests error-handling patterns
Unified API client: Consistent interface across 900+ services
Self-hosted or cloud: Run on your own infrastructure or use Nango Cloud
How to use
Clone the repo and configure your environment (Docker recommended for self-hosting)
Register the APIs you want to integrate with in the Nango dashboard
Use the SDK to trigger OAuth flows and retrieve tokens in your app
Write integration logic in TypeScript, or use Nango's AI assistant to generate sync scripts from a natural-language description of your data flow
Monitor and debug integrations from the dashboard
🔗 View on GitHub | GitHub stars: 11,502
Learning resources
Nango Quickstart Guide - Nango Docs - Official walkthrough covering environment setup, API registration, OAuth flow integration, and writing your first sync script.
4. testcontainers-java
Why this made the cut: The gold standard for spinning up real databases, browsers, and services in your JUnit tests, no mocks required (just Docker).
Why it matters
Mocking databases and services in tests only gets you so far. Testcontainers gives you real, ephemeral instances of PostgreSQL, Redis, Kafka, Selenium, or any Dockerized service, directly in your test suite. Tests become more reliable, and your CI environment matches production more closely.
Key features
First-class JUnit support: Integrates seamlessly with JUnit 4 and JUnit 5
Pre-built modules: Postgres, MySQL, Kafka, Elasticsearch, Selenium, and more
Docker Compose support: Spin up multi-container environments for complex integration tests
Community-driven: 527 contributors and active Hacktoberfest participation
How to use
Add the Testcontainers dependency to your Maven or Gradle project:
<dependency>
<groupId>org.testcontainers</groupId>
<artifactId>testcontainers</artifactId>
<scope>test</scope>
</dependency>Annotate your test class with @Testcontainers and declare container fields:
@Container
static PostgreSQLContainer<?> postgres = new PostgreSQLContainer<>("postgres:15");Testcontainers will start the container before your tests and tear it down afterward
Use postgres.getJdbcUrl() in your test setup to connect to the real database
For multi-service scenarios, use Docker Compose files with the DockerComposeContainer class
🔗 View on GitHub | GitHub stars: 8,709
Learning resources
Scaling Integration Tests: Parallel Spring Tests with JUnit & Testcontainers - JAVAPRO - Conference talk covering parallel integration testing with Testcontainers and JUnit in Spring Boot projects.
5. cua
Why this made the cut: Open-source infrastructure for building, training, and benchmarking AI computer-use agents across macOS, Windows, and Linux.
Why it matters
AI agents that control computers are no longer science fiction, but most implementations are closed and opaque. Cua provides open drivers, cross-OS virtualization (including macOS and Windows sandboxes), and benchmarks for training and evaluation. If you want to build or audit the next generation of desktop automation, this is the starting point.
Key features
Cross-OS support: Drivers for macOS (via Apple Virtualization Framework), Windows (via Windows Sandbox), and Linux
Benchmarking and data generation: Built-in tools for training and evaluating computer-use agents
Fleet orchestration: Scale up agent fleets for parallel evaluation or data collection
Extensible architecture: Designed for researchers and developers to add new drivers and tasks
Desktop automation primitives: Mouse click/move, keyboard input, and screen capture APIs that let agents interact with any desktop application programmatically
How to use
Clone the repo and install dependencies with pip install cua-computer
Set up the appropriate virtualization backend (Apple Virtualization Framework for macOS, Windows Sandbox for Windows)
Run the included benchmarks to verify your environment
Write a basic agent script using the mouse, keyboard, and screen capture APIs, for example, capture a screenshot, locate a UI element, then send a click and keystrokes
Scale up with fleet orchestration to run parallel agent evaluations or data-collection jobs
🔗 View on GitHub | GitHub stars: 21,548
Learning resources
Cua Documentation - Cua - Official docs covering virtualization setup, driver configuration, and building your first computer-use agent.
If you only try one
Start with umami. If you are running any web property, you probably need analytics, and Umami lets you set up a privacy-first, self-hosted dashboard in under an hour. With 38,000+ stars and a mature codebase, it is production-ready and a concrete way to replace a black-box service with something you control. Once you see how painless self-hosting can be, the other projects on this list become a lot less intimidating.
If you are doing Open Source I have a good news for you, I work at CodeRabbit which is an AI review tool and its free for Open Source, please reach out to me on X or LinkedIn or just send an email on [email protected] if you need help on adopting CodeRabbit.
You can visit our portal below to create a new account and connect your repository and start reviewing your code.
