Codemod Grep: Search GitHub Code with AST Patterns
Traditional grep only does text matching. Finding a specific function call pattern means hand-crafting regex, and cross-repository search is a nightmare. Codemod Grep takes a different approach — it understands code syntax.
Powered by ast-grep, you write an AST pattern and search across public GitHub repositories for matching code snippets. Want to find every async/await wrapper around fetch calls? Or a specific combination of React hooks? Just describe the code shape.
In practice, it's especially useful for spotting console.log misused in production code — traditional grep catches mentions in comments too, but AST search returns only actual call sites.
Who it's for: Developers doing code reviews, architecture migrations, or anyone wanting to see how open-source projects solve a particular problem.
So Very Bright: Make Your Logo Genuinely Glow on HDR Screens
HDR displays have made one design problem obvious: no matter how refined your logo looks, it's just another brightness-level color block on an HDR screen. So Very Bright addresses this exact detail.
Upload a logo, select which color regions should glow, download an HDR JPEG. On HDR-capable screens, selected areas shine up to 7.5x brighter than pure white (#FFFFFF), while everything else stays at standard JPEG brightness.
The whole process runs in your browser, free, no registration required. The downloaded HDR JPEG is a standard format — just swap it in for your website or app logo. Worth trying for brands doing visual upgrades.
writing-eval: Local Audit Tool for AI Writing Style
AI-generated text is getting harder to distinguish, but teams have their own standards for "writing style." writing-eval provides a deterministic detection pipeline: extract style features from your reference text, build a profile, then score drafts against it item by item.
Checks cover clarity, readability, sentence rhythm, vocabulary usage, and writing pattern identification. Reports come as Markdown, with each issue annotated with current and target values plus specific editing suggestions.
Runs entirely locally — no API calls, no data uploads. For content teams, this is a repeatable quality gate: run it on a first draft and you'll catch style drift faster than manual review.
Who it's for: Teams running technical blogs, product documentation, or anyone needing to maintain brand voice consistency.