Input/Output nodes of checkpoint (.ckpt) files

[JLenzy]

I have the following files available:

model.ckpt-3668.data-00000-of-00001
model.ckpt-3668.meta
model.ckpt-3668.index

I would like to get the input/output nodes of this graph so I can convert to ONNX. But after building tensorflow with Bazel per the instructions, it seems that it is only expecting a .pb file (instead of a .ckpt meta file). Here are the initial errors it gives:
[libprotobuf ERROR external/com_google_protobuf/src/google/protobuf/wire_format_lite.cc:584] String field 'tensorflow.NodeDef.op' contains invalid UTF-8 data when parsing a protocol buffer. Use the 'bytes' type if you intend to send raw bytes.
[libprotobuf ERROR external/com_google_protobuf/src/google/protobuf/text_format.cc:324] Error parsing text-format tensorflow.GraphDef: 2:1: Interpreting non ascii codepoint 243.
[libprotobuf ERROR external/com_google_protobuf/src/google/protobuf/text_format.cc:324] Error parsing text-format tensorflow.GraphDef: 2:1: Expected identifier, got: ?

My question is, has anyone succesfully read the input/output nodes without a .pb file?