Weekly: Top Trending Open Source Projects (2026-03-24)
March 24, 2026
This week's trending list was defined by two themes: the continued maturation of AI infrastructure, and a wave of developer tools that prioritize speed above all else.
The biggest mover was astral-sh/ruff, the Rust-based Python linter and formatter, which shipped a major release adding full type-checking capabilities. This puts it in direct competition with mypy and pyright — but running 100x faster. The Python community's reaction was electric, with several major projects announcing immediate adoption.
microsoft/vscode had a landmark week: the team shipped an experimental "agent mode" that lets the editor autonomously make multi-file edits based on natural language instructions. The feature is opt-in and clearly labeled as experimental, but the demos circulating on social media drove a massive spike in stars and forks.
ollama/ollama crossed 100,000 GitHub stars, cementing its position as the easiest way to run LLMs locally. The project's growth reflects a broader trend: developers want to experiment with AI without sending data to external APIs or paying per token.
In the infrastructure space, grafana/grafana released a major update to its alerting system with AI-powered anomaly detection built in. The feature uses a local model to establish baselines and flag deviations, requiring no external AI service.
shadcn/ui continued its dominance of the React component space, with the team shipping a new "blocks" system — pre-built page sections (hero, pricing, dashboard) that can be dropped into any project. The quality bar is high and the customizability is complete.
The week's sleeper hit was gleam-lang/gleam, a type-safe functional language that compiles to Erlang and JavaScript. A viral blog post about using it for a high-concurrency API drove a 300% spike in weekly stars. The language's friendly error messages and zero-cost abstractions are winning converts from both Elixir and TypeScript communities.