# Samosa Chat

A local chat application and HTTP gateway for running large language models entirely on a personal machine, with a model catalog for verified downloads and on-demand vision support for images and PDFs.

Samosa is a local model application and HTTP gateway for chatting with language models that run entirely on the user's own hardware, with no cloud account or telemetry. It starts with no models installed, and a browser interface lets the user download, verify and load models from a curated catalog, including options designed to run on a machine with as little as 16 gigabytes of RAM.

Chat images and PDFs are routed through local text and OCR extraction, with an on-demand visual specialist model brought in only when needed rather than kept resident, and inputs are bounded by the machine's current memory and thermal conditions. It includes an internet-search feature that needs no API key, and integrates a bundled local service for portable folder-based memory. Only one primary chat model is kept loaded in memory at a time.

Website: https://github.com/deepanwadhwa/samosa-chat