CGraph

Why we included this project

CGraph is a cross-platform DAG execution engine written in pure C++ with no third-party dependencies. Instead of hand-wiring threads, you subclass a node, implement its run() method, declare the dependencies, and the framework handles the rest: sequential steps, parallel fan-out, conditional branches, loops, even pub/sub messaging between nodes. That deterministic control over what runs in parallel and what must wait is what makes it useful for teams building multi-step pipelines like AI agent chains, data-processing flows, or service orchestration. The Python bindings (pycgraph) let you prototype in Python while keeping the C++ core's performance. The repo also ships worked examples, including a streaming perception pipeline and a pooled HTTP-server pattern, which are worth reading before you design your own graph.