Continuum

A unified runtime for token + tensor execution — one IR, cache-aware scheduling, multi-tier reuse.

C++ core Python bindings Durable checkpoints Semantic reuse
>>>

python_api

High-level public API, sessions, reuse policies, examples, and integration surface.

read the sphinx docs →
{ }

cpp_api

Core runtime internals — IR, interpreter, backend contracts, and the ABI surface.

read the doxygen docs →
$ git

source

Repository, benchmarks, examples, and the research paper behind the runtime.

view on github →