Agentic framework for long-running, creative reasoning tasks. Built for large scale multi-agent reasoning with branching, long-term memory and complex tool calls.
Symbolic-numeric computing framework for dynamical systems. Combines symbolic reasoning with simulations through a LEAN-based proof system, category theoretic circuit compiler, and simulation runtime.
Foundational model for dynamical systems. A transformer-based model trained on Asgard's representation, capable of generating, rewriting and expanding dynamical systems.
Agents for reasoning about dynamical systems. Ties Hugin, Asgard and Mimir together, combining formal reasoning, systems generation and numerical simulations.