From 30f5418d3c250c4cee763f8a582e7f2c89b4cefd Mon Sep 17 00:00:00 2001
From: Enrique Romero <enrique.romero1110@gmail.com>
Date: Wed, 23 Apr 2025 14:01:34 -0400
Subject: [PATCH] Add readme

---
 internal/listeners/README.md | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)
 create mode 100644 internal/listeners/README.md

diff --git a/internal/listeners/README.md b/internal/listeners/README.md
new file mode 100644
index 0000000..66da495
--- /dev/null
+++ b/internal/listeners/README.md
@@ -0,0 +1,25 @@
+# Contacting the WebSocket
+
+When sending information over the webSocket, if it
+is audio make sure to send it as a binary blob. If
+sending over text make sure it's sent as a text blob.
+
+## Starting Speech to Text
+
+We use Vosk for our speech to text service. To start
+the docker container used for vosk run the following
+command
+
+`docker run -d --rm -p 2700:2700 alphacep/kaldi-en:latest`
+
+If you have a cuda capable gpu and would like to use
+it to assist the speech to text process run the following
+
+`docker run -d --rm -p 2700:2700 alphacep/kaldi-en-gpu:latest`
+
+## Contacting Speech to Text
+
+When sending audio over to the go project websocket,
+make sure it is downsampled to 16khz as that is what
+vosk expects. The project itself handles contacting
+vosk directly and sending the audio to it.