new A0 llm endpoint, switch utility calls to non-stream
frdel committed
Feb 3, 2026 at 21:35 UTC
d05224b895baab99ebc2f1cad55ddecd8bf35e69
4 files changed
+21
-12
conf/model_providers.yaml
+1
-1
@@ -21,7 +21,7 @@ chat:
21
name: Agent Zero API
22
litellm_provider: openai
23
kwargs:
24
- api_base: https://api.agent-zero.ai/venice/v1
24
+ api_base: https://llm.agent-zero.ai/v1
25
venice_parameters:
26
include_venice_system_prompt: false
27
anthropic:
python/extensions/message_loop_prompts_after/_50_recall_memories.py
+5
-4
@@ -63,9 +63,9 @@ class RecallMemories(Extension):
63
# get system message and chat history for util llm
64
system = self.agent.read_prompt("memory.memories_query.sys.md")
65
66
- # log query streamed by LLM
67
- async def log_callback(content):
68
- log_item.stream(query=content)
66
+ # # log query streamed by LLM
67
+ # async def log_callback(content):
68
+ # log_item.stream(query=content)
69
70
# call util llm to summarize conversation
71
user_instruction = (
@@ -83,9 +83,10 @@ class RecallMemories(Extension):
83
query = await self.agent.call_utility_model(
84
system=system,
85
message=message,
86
- callback=log_callback,
86
+ # callback=log_callback,
87
)
88
query = query.strip()
89
+ log_item.update(query=query) # no need for streaming here
90
except Exception as e:
91
err = errors.format_error(e)
92
self.agent.context.log.log(
python/extensions/monologue_end/_50_memorize_fragments.py
+7
-4
@@ -44,18 +44,21 @@ class MemorizeMemories(Extension):
44
system = self.agent.read_prompt("memory.memories_sum.sys.md")
45
msgs_text = self.agent.concat_messages(self.agent.history)
46
47
- # log query streamed by LLM
48
- async def log_callback(content):
49
- log_item.stream(content=content)
47
+ # # log query streamed by LLM
48
+ # async def log_callback(content):
49
+ # log_item.stream(content=content)
50
51
# call util llm to find info in history
52
memories_json = await self.agent.call_utility_model(
53
system=system,
54
message=msgs_text,
55
- callback=log_callback,
55
+ # callback=log_callback,
56
background=True,
57
)
58
59
+ # log data < no need for streaming utility messages
60
+ log_item.update(content=memories_json)
61
+
62
# Add validation and error handling for memories_json
63
if not memories_json or not isinstance(memories_json, str):
64
log_item.update(heading="No response from utility model.")
python/extensions/monologue_end/_51_memorize_solutions.py
+8
-3
@@ -41,17 +41,22 @@ class MemorizeSolutions(Extension):
41
msgs_text = self.agent.concat_messages(self.agent.history)
42
43
# log query streamed by LLM
44
- async def log_callback(content):
45
- log_item.stream(content=content)
44
+ # async def log_callback(content):
45
+ # log_item.stream(content=content)
46
47
# call util llm to find solutions in history
48
solutions_json = await self.agent.call_utility_model(
49
system=system,
50
message=msgs_text,
51
- callback=log_callback,
51
+ # callback=log_callback,
52
background=True,
53
)
54
55
+ # log query < no need for streaming utility messages
56
+ log_item.update(content=solutions_json)
57
+
58
+
59
+
60
# Add validation and error handling for solutions_json
61
if not solutions_json or not isinstance(solutions_json, str):
62
log_item.update(heading="No response from utility model.")