cactus-compute

3-Tier hybrid AI router that orchestrates FunctionGemma-270M on-device and Gemini 2.5 Flash Lite in the cloud for 99% function-calling accuracy at 548ms avg latency. Built at the Cactus × Google DeepMind Hackathon.

Golang rewrite of Cactus-Compute needle2 library

OpenAI-compatible HTTP API for Needle 2, Cactus Compute's 45M-parameter tool-calling model. Drop-in for any OpenAI client.

AI-powered emergency sound detection for people with disabilities. On-device FunctionGemma via Cactus Compute. Sub-50ms response. 82% benchmark score.

A SpeechToText application that uses OpenAI's whisper via faster-whisper to transcribe audio and send that information to VRChats textbox system and/or KillFrenzyAvatarText over OSC. Also supports various other methods like OBS via Browsersource and a SteamVR overlay!

A React Native app that runs selective models locally after downloading