From d05224b895baab99ebc2f1cad55ddecd8bf35e69 Mon Sep 17 00:00:00 2001
From: frdel <38891707+frdel@users.noreply.github.com>
Date: Tue, 3 Feb 2026 21:35:05 +0100
Subject: [PATCH] new A0 llm endpoint, switch utility calls to non-stream

---
 conf/model_providers.yaml                             |  2 +-
 .../message_loop_prompts_after/_50_recall_memories.py |  9 +++++----
 .../monologue_end/_50_memorize_fragments.py           | 11 +++++++----
 .../monologue_end/_51_memorize_solutions.py           | 11 ++++++++---
 4 files changed, 21 insertions(+), 12 deletions(-)

diff --git a/conf/model_providers.yaml b/conf/model_providers.yaml
index e6b3c18ec..207f93633 100644
--- a/conf/model_providers.yaml
+++ b/conf/model_providers.yaml
@@ -21,7 +21,7 @@ chat:
     name: Agent Zero API
     litellm_provider: openai
     kwargs:
-      api_base: https://api.agent-zero.ai/venice/v1
+      api_base: https://llm.agent-zero.ai/v1
       venice_parameters:
         include_venice_system_prompt: false
   anthropic:
diff --git a/python/extensions/message_loop_prompts_after/_50_recall_memories.py b/python/extensions/message_loop_prompts_after/_50_recall_memories.py
index aee20175c..9e2beb9f8 100644
--- a/python/extensions/message_loop_prompts_after/_50_recall_memories.py
+++ b/python/extensions/message_loop_prompts_after/_50_recall_memories.py
@@ -63,9 +63,9 @@ class RecallMemories(Extension):
         # get system message and chat history for util llm
         system = self.agent.read_prompt("memory.memories_query.sys.md")
 
-        # log query streamed by LLM
-        async def log_callback(content):
-            log_item.stream(query=content)
+        # # log query streamed by LLM
+        # async def log_callback(content):
+        #     log_item.stream(query=content)
 
         # call util llm to summarize conversation
         user_instruction = (
@@ -83,9 +83,10 @@ class RecallMemories(Extension):
                 query = await self.agent.call_utility_model(
                     system=system,
                     message=message,
-                    callback=log_callback,
+                    # callback=log_callback,
                 )
                 query = query.strip()
+                log_item.update(query=query) # no need for streaming here
             except Exception as e:
                 err = errors.format_error(e)
                 self.agent.context.log.log(
diff --git a/python/extensions/monologue_end/_50_memorize_fragments.py b/python/extensions/monologue_end/_50_memorize_fragments.py
index 3a6420ae7..9060aef2d 100644
--- a/python/extensions/monologue_end/_50_memorize_fragments.py
+++ b/python/extensions/monologue_end/_50_memorize_fragments.py
@@ -44,18 +44,21 @@ class MemorizeMemories(Extension):
         system = self.agent.read_prompt("memory.memories_sum.sys.md")
         msgs_text = self.agent.concat_messages(self.agent.history)
 
-        # log query streamed by LLM
-        async def log_callback(content):
-            log_item.stream(content=content)
+        # # log query streamed by LLM
+        # async def log_callback(content):
+        #     log_item.stream(content=content)
 
         # call util llm to find info in history
         memories_json = await self.agent.call_utility_model(
             system=system,
             message=msgs_text,
-            callback=log_callback,
+            # callback=log_callback,
             background=True,
         )
 
+        # log data < no need for streaming utility messages
+        log_item.update(content=memories_json)
+
         # Add validation and error handling for memories_json
         if not memories_json or not isinstance(memories_json, str):
             log_item.update(heading="No response from utility model.")
diff --git a/python/extensions/monologue_end/_51_memorize_solutions.py b/python/extensions/monologue_end/_51_memorize_solutions.py
index 7189174fa..eec574fdb 100644
--- a/python/extensions/monologue_end/_51_memorize_solutions.py
+++ b/python/extensions/monologue_end/_51_memorize_solutions.py
@@ -41,17 +41,22 @@ class MemorizeSolutions(Extension):
         msgs_text = self.agent.concat_messages(self.agent.history)
 
         # log query streamed by LLM
-        async def log_callback(content):
-            log_item.stream(content=content)
+        # async def log_callback(content):
+        #     log_item.stream(content=content)
 
         # call util llm to find solutions in history
         solutions_json = await self.agent.call_utility_model(
             system=system,
             message=msgs_text,
-            callback=log_callback,
+            # callback=log_callback,
             background=True,
         )
 
+        # log query < no need for streaming utility messages
+        log_item.update(content=solutions_json)
+
+
+
         # Add validation and error handling for solutions_json
         if not solutions_json or not isinstance(solutions_json, str):
             log_item.update(heading="No response from utility model.")