Introducing AI-Powered Snippet Suggestions

< p > Today we're releasing the most requested feature of 2025: AI-Powered Snippet Suggestions. As you write code, Snigo analyzes the context and surfaces relevant snippets from your library — automatically.

< h2 > How It Works < p > Our AI model runs locally on your machine using Apple's MLX framework on macOS and ONNX Runtime on Windows/Linux. Your code never leaves your device — there are no API calls, no cloud processing, and no latency.

< h3 > Context Analysis < p > The model analyzes three signals:

< ol >

Current file context: What language, framework, and patterns are you using?

Recent activity: What snippets have you used recently?

Cursor position: What are you about to write based on the surrounding code?

< h3 > Smart Ranking < p > Suggestions are ranked by relevance, not just similarity.A snippet you use weekly ranks higher than one you saved a year ago and never touched.Snippets from the same project collection get a boost.

< h2 > Privacy First < p > We know developers are rightfully cautious about AI and code privacy.That's why our model runs 100% locally:

< ul >

No code is sent to any server

< li > The model weights are downloaded once and run locally < li > You can disable AI suggestions entirely in settings < li > The feature works completely offline < h2 > Getting Started < p > Update to Snigo v3.2 + and enable AI Suggestions in Settings → AI.The model downloads automatically(approximately 150MB).Suggestions appear as ghost text in the VS Code extension and as a sidebar panel in the desktop app.