Saurav Panda

Run local LLMs like llama, deepseek-distill, kokoro and more inside your browser

WASM-first LLM inference engine in pure Rust. WebGPU-accelerated, GGUF-native — run LLMs in the browser with zero server costs.

Safe git worktree cleanup for coding agents. Finds stale Claude Code worktrees across your machine, classifies them safe/review/protected, and reclaims the disk space they leave behind — without ever deleting unpushed work.

Simple API for deploying whisper on your own server

Rust-native AOT compiler for small LLMs. Turns GGUF weights into optimized self-contained binaries — CPU SIMD, Metal, WASM, no interpreter.

An AI agent to review your developer docs and give critical feedback!