# C++ relay frontend

**URL:** <https://discuss.tvm.apache.org/t/c-relay-frontend/18917>\
**Category:** Questions\
**Created:** [February 2, 2026, 2:38pm UTC](https://discuss.tvm.apache.org/t/c-relay-frontend/18917 "2026-02-02T14:38:08Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![binzhang1008](https://discuss.tvm.apache.org/user_avatar/discuss.tvm.apache.org/binzhang1008/32/6148_2.png) [@binzhang1008](https://discuss.tvm.apache.org/u/binzhang1008)\
**Post date:** [February 2, 2026, 2:38pm UTC](https://discuss.tvm.apache.org/t/c-relay-frontend/18917/1 "2026-02-02T14:38:08Z")

</div>

Currently, it seems that in TVM, the ONNX input is first received through the Python interface `relay.frontend.from_onnx`, and then the model is compiled using `relay.build`. However, my environment is purely C++, and I haven’t found the relevant frontend interfaces in the `src` directory. I would like to ask if TVM supports a purely C++ workflow, including `from_onnx`, generating `IRModule`, and building the model. If so, which branch should I use? If not, what work do I need to do to implement these functionalities?