>>103299191
Piper. The OCR part is more complicated, but assuming you can figure that out you can just replace the echo test below with your OCR output. Also change the path to your voice model. You can pipe any text to it this way, henceforth the name.
echo 'This is a test' | piper -m "your_voice_model.onnx" --output-raw | aplay -r 22050 -f S16_LE -t raw -
https://github.com/rhasspy/piper