Needle
Foundation model optimized for edge devices with tool calling and embeddings.
0 members list Needle, unchanged over the last 12 weeks, data as of 9 November 2026
Members who list Needle in their Stack, by week.
Members who list Needle
Be the first to add Needle
Nobody on Stackness runs Needle yet. Add it to your stack and you're the first member running it.
Homebrew analytics installs: 25 installs (2026) to 24 installs (2026)
Imported from Homebrew analytics, last reading Oct 2026
About Needle
Needle is a foundation model developed by Cactus Compute designed to run on edge devices with minimal memory footprint. It enables tool calling and embeddings on mobile devices, wearables, smart home systems, robots, automotive systems, and other resource-constrained hardware. The model achieves this through sub-2-bit quantization, allowing large models to fit into devices with minimal processing power and storage. Needle can operate entirely offline or use hybrid routing to fall back to cloud resources when needed, based on confidence scoring.
The model family includes variants ranging from 8 to 29 MB in size that perform on-device inference without requiring GPU acceleration or internet connectivity. Needle is built to handle agent-based tasks, voice processing, and vision capabilities across diverse hardware platforms from microcontrollers to laptops and game consoles. The tool is available as open source and can be accessed through Hugging Face for download and integration into applications.
Checked 9 November 2026.
Nobody on Stackness runs Needle yet.
Sign up and add it