From 811a40f2cc88c0a8b136093a78bbdf8fbe355e72 Mon Sep 17 00:00:00 2001
From: Adolfo Reyna <aeroreyna@gmail.com>
Date: Wed, 12 Aug 2026 10:38:23 -0400
Subject: [PATCH] fix: isolate reasoning chunks from assistant spoken reply
 history and Web UI final reply

---
 hermes_llm.py | 18 +++++++++++++-----
 1 file changed, 13 insertions(+), 5 deletions(-)

diff --git a/hermes_llm.py b/hermes_llm.py
index 896c3bf..f0cb581 100644
--- a/hermes_llm.py
+++ b/hermes_llm.py
@@ -451,6 +451,7 @@ class HermesLLM(FrameProcessor):
         if not suppress_output:
             await self.push_frame(LLMFullResponseStartFrame())
         chunks: list[str] = []
+        spoken_chunks: list[str] = []
 
         if self._keep_open:
             await self._ensure_persistent_proc()
@@ -470,9 +471,9 @@ class HermesLLM(FrameProcessor):
             pass
 
         if server_ok or self._use_server:
-            await self._run_turn_server(utterance, chunks)
+            await self._run_turn_server(utterance, chunks, spoken_chunks)
         else:
-            await self._run_turn_cli(utterance, chunks)
+            await self._run_turn_cli(utterance, chunks, spoken_chunks)
 
         t1 = time.perf_counter()
         total_ms = int((t1 - t0) * 1000)
@@ -491,7 +492,7 @@ class HermesLLM(FrameProcessor):
         except Exception:
             pass
 
-        full_reply = _clean_spoken_text(" ".join(chunks))
+        full_reply = _clean_spoken_text(" ".join(spoken_chunks)) if 'spoken_chunks' in locals() and spoken_chunks else _clean_spoken_text(" ".join(chunks))
         if full_reply:
             self._history.append({"role": "assistant", "content": full_reply})
             logger.info(f"Hermes LLM ({self._model or 'default'}): {full_reply}")
@@ -504,8 +505,10 @@ class HermesLLM(FrameProcessor):
                 if self._on_reply:
                     self._on_reply(full_reply)
 
-    async def _run_turn_server(self, utterance: str, chunks: list[str]):
+    async def _run_turn_server(self, utterance: str, chunks: list[str], spoken_chunks: list[str] | None = None):
         """Run turn via Hermes OpenAI-compatible Gateway API using SSE streaming (stream: true)."""
+        if spoken_chunks is None:
+            spoken_chunks = chunks
         try:
             ok, _ = await check_hermes_server_active(self._port)
             if not ok:
@@ -616,6 +619,7 @@ class HermesLLM(FrameProcessor):
                                                     if not cleaned_sent.endswith((".", "!", "?")):
                                                         cleaned_sent += "."
                                                     chunks.append(cleaned_sent)
+                                                    spoken_chunks.append(cleaned_sent)
                                                     await self.push_frame(LLMTextFrame(cleaned_sent))
                                                     try:
                                                         import web_server
@@ -639,6 +643,7 @@ class HermesLLM(FrameProcessor):
                             if not cleaned_rem.endswith((".", "!", "?")):
                                 cleaned_rem += "."
                             chunks.append(cleaned_rem)
+                            spoken_chunks.append(cleaned_rem)
                             await self.push_frame(LLMTextFrame(cleaned_rem))
                             try:
                                 import web_server
@@ -666,8 +671,10 @@ class HermesLLM(FrameProcessor):
                 pass
             await self._run_turn_cli(utterance, chunks)
 
-    async def _run_turn_cli(self, utterance: str, chunks: list[str]):
+    async def _run_turn_cli(self, utterance: str, chunks: list[str], spoken_chunks: list[str] | None = None):
         """Run turn via Hermes CLI using persistent session tracking."""
+        if spoken_chunks is None:
+            spoken_chunks = chunks
         cmd = [self._cli_path, "chat", "-q", utterance, "-Q", "--source", "voice", "--reasoning", "none"]
         if self._session_id:
             cmd.extend(["-r", self._session_id])
@@ -735,6 +742,7 @@ class HermesLLM(FrameProcessor):
                 cleaned = _clean_spoken_text(text_line)
                 if cleaned:
                     chunks.append(cleaned)
+                    spoken_chunks.append(cleaned)
                     await self.push_frame(LLMTextFrame(cleaned))
                     try:
                         import web_server