PyC Compiler Runtime
Deterministic compiler rails for AI workloads
PyC is a C/CMake compiler runtime project focused on reproducible CI, memory-backpressure control, and policy-driven execution decisions. Use prebuilt binaries for fast onboarding, or build from source for full control.
Latest release page: open release details
Loading release metadata...
What this project does
- Stable core targets:
pyc,pyc_core,pyc_foundation - AI runtime optimization modes:
balanced,memory_first,utilization_first - Operational rails: guarded switching, rollback triggers, benchmark gates, and source coverage enforcement
Download binaries
Pick your platform. These links always target the latest release assets.
- Linux: checking latest artifact...
- macOS: checking latest artifact...
- Windows: checking latest artifact...
Each package includes:
bin/pyc(orpyc.exe)lib/static libraries (pyc_core,pyc_foundation)include/headers
Quick smoke command after extraction:
./bin/pyc
# Windows: .\bin\pyc.exe
Build from source
If you want full control or development workflow integration:
cmake -S . -B build -D PYC_BUILD_COMPILER_NEXT=ON -D PYC_BUILD_COMPILER_NEXT_TESTS=ON
cmake --build build --parallel
ctest --test-dir build -C Release --output-on-failure
Stable smoke:
./build/pyc
# Windows: .\build\Release\pyc.exe
Integrate in AI runtime jobs
Put PyC under your job runtime to apply dynamic policy control during execution windows. It is meant to stay out of your way: deterministic defaults, explicit modes, and rollback rails when pressure rises.
- Warmup:
utilization_first - Steady state:
balanced - Pressure event:
memory_first
Detailed integration guide: runtime integration spec.