The Assistant That Refuses to Guess
A deterministic answer layer in front of a local LLM: it serves what it can prove from a knowledge graph and refuses the rest. The refusals turned out to be the interesting part.
Read the article →AME LABS
Artificial Minds Engineering
AME Labs is an independent AI engineering lab in Paris. We are building AME-1, an artificial mind on a neuro-symbolic inference architecture.
AME-1, the framework
A character with a governed mind, living on your hardware: persistent identity, real memory, honesty about its own limits, ethics in the pipeline — and the ability to act, inside an envelope it cannot leave.
Meet the framework →The architecture
An extensible modular monolith: one Rust binary carrying a specified 21-module mind, every dependency behind a swappable adapter, every new capability an extension package.
Explore the architecture →