ELI v2.0
A private AI assistant and local runtime that runs entirely on the user's own hardware: it holds conversations by voice or text, controls the desktop, reads the screen, writes code, and builds a private model of the user over time.
ELI is a local, private AI assistant and cognitive runtime that runs entirely on the user's own machine, with no cloud account. It holds conversations by voice or text, controls the desktop, reads the screen and documents, writes and fixes code, and builds up a private picture of the user the more it is used. It enforces offline operation at the network socket rather than by convention, and works with any local GGUF model, tuning itself to the hardware it finds.
Under its conversational surface it runs a 12-stage reasoning pipeline with a fleet of specialist agents on a DAG orchestrator, layered memory built from SQLite, a FAISS vector index and a knowledge graph, local voice and vision, and its own smart-home server, all reachable through one typed or spoken interface with 228 capabilities. A remote view lets a phone reach it over the user's own network.
It is written in Python. Version 2 is a public, actively developed line, while a newer version is being developed privately.