TensorFlow Serving (TF Serving) is a tool to run TensorFlow models online in large production settings using a RPC or REST API. TensorFlow Decision Forests (TF-DF) is supported natively by TF Serving >=2.11.
TF-DF models are directly compatible with TF Serving. Yggdrasil models can be used with TF Serving after being converted first.
Check our TF Serving + TF-DF tutorial on tensorflow.org for more details.
Old versions of TF Serving (e.g. TF Serving 2.8-2.10) are compatible with TF-DF. However, they require to be re-compiled with TF-DF support (instructions).