langwatch-scenario 0.7.7__py3-none-any.whl → 0.7.9__py3-none-any.whl

{langwatch_scenario-0.7.7.dist-info → langwatch_scenario-0.7.9.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: langwatch-scenario
-Version: 0.7.7
+Version: 0.7.9
 Summary: The end-to-end agent testing library
 Author-email: LangWatch Team <support@langwatch.ai>
 License: MIT
@@ -457,7 +457,7 @@ This will cache any function call you decorate when running the tests and make t
 While optional, we strongly recommend setting stable identifiers for your scenarios, sets, and batches for better organization and tracking in LangWatch.
 
 - **set_id**: Groups related scenarios into a test suite. This corresponds to the "Simulation Set" in the UI.
-- **batch_run_id**: Groups all scenarios that were run together in a single execution (e.g., a single CI job). This is automatically generated but can be overridden.
+- **SCENARIO_BATCH_RUN_ID**: Env variable that groups all scenarios that were run together in a single execution (e.g., a single CI job). This is automatically generated but can be overridden.
 
 ```python
 import os

{langwatch_scenario-0.7.7.dist-info → langwatch_scenario-0.7.9.dist-info}/RECORD

@@ -2,20 +2,21 @@ scenario/__init__.py,sha256=4WO8TjY8Lc0NhYL7b9LvaB1xCBqwUkLuI0uIA6PQP6c,4223
 scenario/_error_messages.py,sha256=QVFSbhzsVNGz2GOBOaoQFW6w6AOyZCWLTt0ySWPfnGw,3882
 scenario/agent_adapter.py,sha256=PoY2KQqYuqzIIb3-nhIU-MPXwHJc1vmwdweMy7ut-hk,4255
 scenario/cache.py,sha256=J6s6Sia_Ce6TrnsInlhfxm6SF8tygo3sH-_cQCRX1WA,6213
-scenario/judge_agent.py,sha256=7NsgeMu6wRMjU_HYTCFqkLma6H2AJuEkw9hJkt11190,16211
-scenario/pytest_plugin.py,sha256=DGrpgB6e71eq8QXWWxwLjAKNhiyYyzfzZ0L5Ax8iEmo,11317
-scenario/scenario_executor.py,sha256=2ZPy2cywwEMIbUfBP1jHN__Ffjf5WGB144MX2SNr5IM,33101
+scenario/judge_agent.py,sha256=TSwykEWhoBA9F__sUsSuUMpu7pOkT1lIJo8YlEj2eiA,16759
+scenario/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+scenario/pytest_plugin.py,sha256=wRCuGD9uwrrLt2fY15zK6mnmY9W_dO_m0WalPJYE5II,11491
+scenario/scenario_executor.py,sha256=_GRpFpw_WtgtaGpxWh0A0HNNf-aU78PdIiVdgEFm9MY,33136
 scenario/scenario_state.py,sha256=LWGqEQN-Yz0DIiC-TyMRHd-9rEiuBVUHKllMmKv-qGg,7029
 scenario/script.py,sha256=A0N5pP0l4FFn1xdKc78U_wkwWhEWH3EFeU_LRDtNyEI,12241
 scenario/types.py,sha256=qH5KFzJBDG1fEJB_qFRVtL3EZulxq3G1mztYczIzIAY,9613
-scenario/user_simulator_agent.py,sha256=UJ75xhqHwoi8-3JkR1AsHDzpHM2Lx-aDSTJ1gnq_SXc,9101
+scenario/user_simulator_agent.py,sha256=kqnSd4_gytzEwtkc06r58UdE1EycZBzejRPzfORDjdo,9619
 scenario/_events/__init__.py,sha256=4cj6H9zuXzvWhT2P2JNdjWzeF1PUepTjqIDw85Vid9s,1500
-scenario/_events/event_alert_message_logger.py,sha256=K0Pu76Gd36lGEEYh8e8r7NMt7J-OQhbw0cZmiwutCOE,3591
-scenario/_events/event_bus.py,sha256=KFN0OxAQIQXIk_tVrorDoN_YLKVK9dos5SXFALstHgE,9809
-scenario/_events/event_reporter.py,sha256=4uND_kdPBXe-aUWCdSj4BLrMA33TDnbZzokAEOU3_08,3771
+scenario/_events/event_alert_message_logger.py,sha256=n2W3uT8y4x6KKL3H9Ez6CfzJOFlvOfvjDKsdhHUJkxs,2787
+scenario/_events/event_bus.py,sha256=IsKNsClF1JFYj728EcxX1hw_KbfDkfJq3Y2Kv4h94n4,9871
+scenario/_events/event_reporter.py,sha256=-6NNbBMy_FYr1O-1FuZ6eIUnLuI8NGRMUr0pybLJrCI,3873
 scenario/_events/events.py,sha256=UtEGY-_1B0LrwpgsNKgrvJBZhRtxuj3K_i6ZBfF7E4Q,6387
 scenario/_events/messages.py,sha256=quwP2OkeaGasNOoaV8GUeosZVKc5XDsde08T0xx_YQo,2297
-scenario/_events/utils.py,sha256=SproqiwjhLWAW7p82EirCgawpxAo0ksW1pBB4mKkcEs,3436
+scenario/_events/utils.py,sha256=KKqWFGkj4XtofKxM2yi-DBhBQp8wQOdls48iPHGCmUY,3473
 scenario/_generated/langwatch_api_client/README.md,sha256=Az5f2L4ChOnG_ZtrdBagzRVgeTCtBkbD_S5cIeAry2o,5424
 scenario/_generated/langwatch_api_client/pyproject.toml,sha256=Z8wxuGp4H9BJYVVJB8diW7rRU9XYxtPfw9mU4_wq4cA,560
 scenario/_generated/langwatch_api_client/lang_watch_api_client/__init__.py,sha256=vVrn17y-3l3fOqeJk8aN3GlStRm2fo0f313l_0LtJNs,368
@@ -226,16 +227,16 @@ scenario/_generated/langwatch_api_client/lang_watch_api_client/models/search_req
 scenario/_generated/langwatch_api_client/lang_watch_api_client/models/search_response.py,sha256=zDYmJ8bFBSJyF9D3cEn_ffrey-ITIfwr-_7eu72zLyk,2832
 scenario/_generated/langwatch_api_client/lang_watch_api_client/models/timestamps.py,sha256=-nRKUPZTAJQNxiKz128xF7DKgZNbFo4G3mr5xNXrkaw,2173
 scenario/_generated/langwatch_api_client/lang_watch_api_client/models/trace.py,sha256=K9Lc_EQOrJ2dqMXx9EpiUXReT1_uYF7WRfYyhlfbi3I,7537
-scenario/_utils/__init__.py,sha256=ptNVzmjhypznnozdNIiuBDHZ0NLqtp7xhio9kEDovWQ,1311
-scenario/_utils/ids.py,sha256=v3JS8J7vrFuubK5bXJviU-BVZoLGWINCN1hUyAO9NZw,2074
+scenario/_utils/__init__.py,sha256=xPVjLXnHTTq9fuRFh5lsMvwtIpEeJ3jy1vf5yTUMPsc,1313
+scenario/_utils/ids.py,sha256=W4tVMCf9ky0KLTDA_qOfErNhb4tCmxwa8zEuo1K1ZuY,2071
 scenario/_utils/message_conversion.py,sha256=AWHn31E7J0mz9sBXWruVVAgtsrJz1R_xEf-dGbX6jjs,3636
 scenario/_utils/utils.py,sha256=msQgUWaLh3U9jIIHmxkEbOaklga63AF0KJzsaKa_mZc,14008
 scenario/config/__init__.py,sha256=b2X_bqkIrd7jZY9dRrXk2wOqoPe87Nl_SRGuZhlolxA,1123
 scenario/config/langwatch.py,sha256=ijWchFbUsLbQooAZmwyTw4rxfRLQseZ1GoVSiPPbzpw,1677
-scenario/config/model.py,sha256=Ve49S2FyzUifXJ-SAyKPiNtVqs8BfsYbODu_M5y0c8Y,1155
-scenario/config/scenario.py,sha256=tVVnsUgG6Z0hYZiTDX-GGZz8l8co1HhyTqJUJNPinBk,5184
-langwatch_scenario-0.7.7.dist-info/METADATA,sha256=L7h0kgOaIij6MYVCac0EqPu8ODkZNKxDeIrHCSJg2l4,20003
-langwatch_scenario-0.7.7.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-langwatch_scenario-0.7.7.dist-info/entry_points.txt,sha256=WlEnJ_gku0i18bIa3DSuGqXRX-QDQLe_s0YmRzK45TI,45
-langwatch_scenario-0.7.7.dist-info/top_level.txt,sha256=45Mn28aedJsetnBMB5xSmrJ-yo701QLH89Zlz4r1clE,9
-langwatch_scenario-0.7.7.dist-info/RECORD,,
+scenario/config/model.py,sha256=T4HYA79CW1NxXDkFlyftYR6JzZcowbtIx0H-ijxRyfg,1297
+scenario/config/scenario.py,sha256=6jrtcm0Fo7FpxQta7QIKdGMgl7cXrn374Inzx29hRuk,5406
+langwatch_scenario-0.7.9.dist-info/METADATA,sha256=0s-yAn8iE1N-5dbqugYFpSl8btZrTyyDgWQDat8szxI,20030
+langwatch_scenario-0.7.9.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+langwatch_scenario-0.7.9.dist-info/entry_points.txt,sha256=WlEnJ_gku0i18bIa3DSuGqXRX-QDQLe_s0YmRzK45TI,45
+langwatch_scenario-0.7.9.dist-info/top_level.txt,sha256=45Mn28aedJsetnBMB5xSmrJ-yo701QLH89Zlz4r1clE,9
+langwatch_scenario-0.7.9.dist-info/RECORD,,

scenario/_events/event_alert_message_logger.py

@@ -1,5 +1,8 @@
 import os
+import webbrowser
 from typing import Set
+
+from ..config.scenario import ScenarioConfig
 from .._utils.ids import get_batch_run_id
 
 
@@ -49,35 +52,13 @@ class EventAlertMessageLogger:
 
         if not os.getenv("LANGWATCH_API_KEY"):
             print(f"\n{separator}")
-            print("🚀  LangWatch Simulation Reporting")
+            print("🎭  Running Scenario Tests")
             print(f"{separator}")
-            print("➡️  API key not configured")
+            print("➡️  LangWatch API key not configured")
             print("   Simulations will only output final results")
             print("")
             print("💡 To visualize conversations in real time:")
             print("   • Set LANGWATCH_API_KEY environment variable")
-            print("   • Or configure apiKey in scenario.config.js")
-            print("")
-            print(f"📦 Batch Run ID: {batch_run_id}")
-            print("")
-            print("🔇 To disable these messages:")
-            print("   • Set SCENARIO_DISABLE_SIMULATION_REPORT_INFO=true")
-            print(f"{separator}\n")
-        else:
-            endpoint = os.getenv("LANGWATCH_ENDPOINT", "https://app.langwatch.ai")
-            api_key = os.getenv("LANGWATCH_API_KEY", "")
-
-            print(f"\n{separator}")
-            print("🚀  LangWatch Simulation Reporting")
-            print(f"{separator}")
-            print("✅ Simulation reporting enabled")
-            print(f"   Endpoint: {endpoint}")
-            print(f"   API Key: {'Configured' if api_key else 'Not configured'}")
-            print("")
-            print(f"📦 Batch Run ID: {batch_run_id}")
-            print("")
-            print("🔇 To disable these messages:")
-            print("   • Set SCENARIO_DISABLE_SIMULATION_REPORT_INFO=true")
             print(f"{separator}\n")
 
     def _display_watch_message(self, set_url: str) -> None:
@@ -86,10 +67,15 @@ class EventAlertMessageLogger:
         batch_url = f"{set_url}/{get_batch_run_id()}"
 
         print(f"\n{separator}")
-        print("👀 Watch Your Simulation Live")
+        print("🎭  Running Scenario Tests")
         print(f"{separator}")
-        print("🌐 Open in your browser:")
-        print(f"   Scenario Set: {set_url}")
-        print(f"   Batch Run: {batch_url}")
-        print("")
+        print(f"Follow it live: {batch_url}")
         print(f"{separator}\n")
+
+        config = ScenarioConfig.default_config
+        if config and not config.headless:
+            # Open the URL in the default browser (cross-platform)
+            try:
+                webbrowser.open(batch_url)
+            except Exception:
+                pass

scenario/_events/event_bus.py

@@ -3,6 +3,7 @@ from typing import Optional, Any, Dict
 from .events import ScenarioEvent
 from .event_reporter import EventReporter
 from .event_alert_message_logger import EventAlertMessageLogger
+from ..config.scenario import ScenarioConfig
 
 import asyncio
 import queue
@@ -35,7 +36,9 @@ class ScenarioEventBus:
     """
 
     def __init__(
-        self, event_reporter: Optional[EventReporter] = None, max_retries: int = 3
+        self,
+        event_reporter: Optional[EventReporter] = None,
+        max_retries: int = 3,
     ):
         """
         Initialize the event bus with optional event reporter and retry configuration.

scenario/_events/event_reporter.py

@@ -3,7 +3,7 @@ import httpx
 from typing import Optional, Dict, Any
 from .events import ScenarioEvent
 from .event_alert_message_logger import EventAlertMessageLogger
-from scenario.config import LangWatchSettings
+from scenario.config import LangWatchSettings, ScenarioConfig
 
 
 class EventReporter:
@@ -26,7 +26,11 @@ class EventReporter:
         reporter = EventReporter(api_key="your-api-key")
     """
 
-    def __init__(self, endpoint: Optional[str] = None, api_key: Optional[str] = None):
+    def __init__(
+        self,
+        endpoint: Optional[str] = None,
+        api_key: Optional[str] = None,
+    ):
         # Load settings from environment variables
         langwatch_settings = LangWatchSettings()
 
@@ -69,6 +73,7 @@ class EventReporter:
                         "Content-Type": "application/json",
                         "X-Auth-Token": self.api_key,
                     },
+                    timeout=httpx.Timeout(30.0),
                 )
                 self.logger.info(
                     f"[{event_type}] POST response status: {response.status_code} ({event.scenario_run_id})"
@@ -92,7 +97,7 @@ class EventReporter:
                     )
         except Exception as error:
             self.logger.error(
-                f"[{event_type}] Event POST error: {error}, event={event}, endpoint={self.endpoint}"
+                f"[{event_type}] Event POST error: {repr(error)}, event={event}, endpoint={self.endpoint}"
             )
 
         return result

scenario/_events/utils.py

@@ -10,7 +10,7 @@ from .messages import (
     FunctionCall,
 )
 from typing import List
-import uuid
+from pksuid import PKSUID
 
 def convert_messages_to_api_client_messages(messages: list[ChatCompletionMessageParam]) -> list[MessageType]:
     """
@@ -33,7 +33,7 @@ def convert_messages_to_api_client_messages(messages: list[ChatCompletionMessage
 
     for i, message in enumerate(messages):
         # Generate unique ID for each message
-        message_id = message.get("id") or str(uuid.uuid4())
+        message_id = message.get("id") or str(PKSUID('scenariomsg'))
 
         role = message.get("role")
         content = message.get("content")
@@ -54,7 +54,7 @@ def convert_messages_to_api_client_messages(messages: list[ChatCompletionMessage
             if tool_calls:
                 for tool_call in tool_calls:
                     api_tool_calls.append(ToolCall(
-                        id=tool_call.get("id", str(uuid.uuid4())),
+                        id=tool_call.get("id", str(PKSUID('scenariotoolcall'))),
                         type_="function",
                         function=FunctionCall(
                             name=tool_call["function"].get("name", "unknown"),

scenario/_utils/__init__.py

@@ -14,7 +14,7 @@ from .ids import (
     generate_scenario_id,
     generate_thread_id,
     generate_message_id,
-    safe_parse_uuid,
+    safe_parse_ksuid,
 )
 from .utils import (
     SerializableAndPydanticEncoder,
@@ -34,7 +34,7 @@ __all__ = [
     "generate_scenario_id",
     "generate_thread_id",
     "generate_message_id",
-    "safe_parse_uuid",
+    "safe_parse_ksuid",
     "SerializableAndPydanticEncoder",
     "SerializableWithStringFallback",
     "print_openai_messages",

scenario/_utils/ids.py

@@ -7,7 +7,7 @@ and scenario tracking.
 """
 
 import os
-import uuid
+from pksuid import PKSUID
 
 
 def generate_thread_id() -> str:
@@ -17,7 +17,7 @@ def generate_thread_id() -> str:
     Returns:
         str: A new thread ID.
     """
-    return f"thread_{uuid.uuid4()}"
+    return f"{PKSUID('scenariothread')}"
 
 
 def generate_scenario_run_id() -> str:
@@ -27,7 +27,7 @@ def generate_scenario_run_id() -> str:
     Returns:
         str: A new scenario run ID.
     """
-    return f"scenariorun_{uuid.uuid4()}"
+    return f"{PKSUID('scenariorun')}"
 
 
 def generate_scenario_id() -> str:
@@ -37,7 +37,7 @@ def generate_scenario_id() -> str:
     Returns:
         str: A new scenario ID.
     """
-    return f"scenario_{uuid.uuid4()}"
+    return f"{PKSUID('scenario')}"
 
 
 def get_batch_run_id() -> str:
@@ -52,7 +52,7 @@ def get_batch_run_id() -> str:
     batch_run_id = os.environ.get("SCENARIO_BATCH_RUN_ID")
     if not batch_run_id:
         # Generate new batch ID if not set
-        batch_run_id = f"scenariobatchrun_{uuid.uuid4()}"
+        batch_run_id = f"{PKSUID('scenariobatch')}"
         os.environ["SCENARIO_BATCH_RUN_ID"] = batch_run_id
 
     return batch_run_id
@@ -65,23 +65,23 @@ def generate_message_id() -> str:
     Returns:
         str: A new message ID.
     """
-    return f"scenariomsg_{uuid.uuid4()}"
+    return f"{PKSUID('scenariomsg')}"
 
 
-def safe_parse_uuid(id_str: str) -> bool:
+def safe_parse_ksuid(id_str: str) -> bool:
     """
-    Safely parses a UUID string.
+    Safely parses a Ksuid string.
 
     Args:
-        id_str: The UUID string to parse.
+        id_str: The Ksuid string to parse.
 
     Returns:
-        bool: True if the UUID string is valid, false otherwise.
+        bool: True if the Ksuid string is valid, false otherwise.
     """
     try:
-        uuid.UUID(id_str)
+        PKSUID.parse(id_str)
         return True
-    except (ValueError, TypeError):
+    except Exception:
         return False
 
 

scenario/config/model.py

@@ -17,7 +17,8 @@ class ModelConfig(BaseModel):
     for use with user simulator and judge agents in the Scenario framework.
 
     Attributes:
-        model: The model identifier (e.g., "openai/gpt-4.1-mini", "anthropic/claude-3-sonnet")
+        model: The model identifier (e.g., "openai/gpt-4.1", "anthropic/claude-3-sonnet")
+        api_base: Optional base URL where the model is hosted
         api_key: Optional API key for the model provider
         temperature: Sampling temperature for response generation (0.0 = deterministic, 1.0 = creative)
         max_tokens: Maximum number of tokens to generate in responses
@@ -26,6 +27,7 @@ class ModelConfig(BaseModel):
         ```
         model_config = ModelConfig(
             model="openai/gpt-4.1",
+            api_base="https://api.openai.com/v1",
             api_key="your-api-key",
             temperature=0.1,
             max_tokens=1000
@@ -34,6 +36,7 @@ class ModelConfig(BaseModel):
     """
 
     model: str
+    api_base: Optional[str] = None
     api_key: Optional[str] = None
     temperature: float = 0.0
     max_tokens: Optional[int] = None

scenario/config/scenario.py

@@ -5,6 +5,7 @@ This module provides the main configuration class for customizing the behavior
 of the Scenario testing framework, including execution parameters and debugging options.
 """
 
+import os
 from typing import Optional, Union, ClassVar
 from pydantic import BaseModel
 
@@ -53,6 +54,11 @@ class ScenarioConfig(BaseModel):
     verbose: Optional[Union[bool, int]] = True
     cache_key: Optional[str] = None
     debug: Optional[bool] = False
+    headless: Optional[bool] = os.getenv("SCENARIO_HEADLESS", "false").lower() not in [
+        "false",
+        "0",
+        "",
+    ]
 
     default_config: ClassVar[Optional["ScenarioConfig"]] = None
 
@@ -64,6 +70,7 @@ class ScenarioConfig(BaseModel):
         verbose: Optional[Union[bool, int]] = None,
         cache_key: Optional[str] = None,
         debug: Optional[bool] = None,
+        headless: Optional[bool] = None,
     ) -> None:
         """
         Set global configuration settings for all scenario executions.
@@ -107,6 +114,7 @@ class ScenarioConfig(BaseModel):
                 verbose=verbose,
                 cache_key=cache_key,
                 debug=debug,
+                headless=headless,
             )
         )
 

scenario/judge_agent.py

@@ -41,6 +41,7 @@ class JudgeAgent(AgentAdapter):
     Attributes:
         role: Always AgentRole.JUDGE for judge agents
         model: LLM model identifier to use for evaluation
+        api_base: Optional base URL where the model is hosted
         api_key: Optional API key for the model provider
         temperature: Sampling temperature for evaluation consistency
         max_tokens: Maximum tokens for judge reasoning
@@ -97,6 +98,7 @@ class JudgeAgent(AgentAdapter):
     role = AgentRole.JUDGE
 
     model: str
+    api_base: Optional[str]
     api_key: Optional[str]
     temperature: float
     max_tokens: Optional[int]
@@ -108,6 +110,7 @@ class JudgeAgent(AgentAdapter):
         *,
         criteria: Optional[List[str]] = None,
         model: Optional[str] = None,
+        api_base: Optional[str] = None,
         api_key: Optional[str] = None,
         temperature: float = 0.0,
         max_tokens: Optional[int] = None,
@@ -122,6 +125,8 @@ class JudgeAgent(AgentAdapter):
                      and negative constraints ("Agent should not provide personal information").
             model: LLM model identifier (e.g., "openai/gpt-4.1").
                    If not provided, uses the default model from global configuration.
+            api_base: Optional base URL where the model is hosted. If not provided,
+                      uses the base URL from global configuration.
             api_key: API key for the model provider. If not provided,
                      uses the key from global configuration or environment.
             temperature: Sampling temperature for evaluation (0.0-1.0).
@@ -156,6 +161,7 @@ class JudgeAgent(AgentAdapter):
         """
         # Override the default system prompt for the judge agent
         self.criteria = criteria or []
+        self.api_base = api_base
         self.api_key = api_key
         self.temperature = temperature
         self.max_tokens = max_tokens
@@ -172,6 +178,9 @@ class JudgeAgent(AgentAdapter):
             ScenarioConfig.default_config.default_model, ModelConfig
         ):
             self.model = model or ScenarioConfig.default_config.default_model.model
+            self.api_base = (
+                api_base or ScenarioConfig.default_config.default_model.api_base
+            )
             self.api_key = (
                 api_key or ScenarioConfig.default_config.default_model.api_key
             )
@@ -351,6 +360,8 @@ if you don't have enough information to make a verdict, say inconclusive with ma
                 model=self.model,
                 messages=messages,
                 temperature=self.temperature,
+                api_key=self.api_key,
+                api_base=self.api_base,
                 max_tokens=self.max_tokens,
                 tools=tools,
                 tool_choice=(
@@ -387,7 +398,7 @@ if you don't have enough information to make a verdict, say inconclusive with ma
                         failed_criteria = [
                             self.criteria[idx]
                             for idx, criterion in enumerate(criteria.values())
-                            if criterion == False
+                            if criterion == False or criterion == "inconclusive"
                         ]
 
                         # Return the appropriate ScenarioResult based on the verdict

scenario/pytest_plugin.py

@@ -199,6 +199,8 @@ class ScenarioReporter:
 # Store the original run method
 original_run = ScenarioExecutor.run
 
+def pytest_addoption(parser):
+    parser.addoption("--headless", action="store_true")
 
 @pytest.hookimpl(trylast=True)
 def pytest_configure(config):
@@ -240,6 +242,9 @@ def pytest_configure(config):
         print(colored("\nScenario debug mode enabled (--debug).", "yellow"))
         ScenarioConfig.configure(verbose=True, debug=True)
 
+    if config.getoption("--headless"):
+        ScenarioConfig.configure(headless=True)
+
     # Create a global reporter instance
     config._scenario_reporter = ScenarioReporter()
 

scenario/scenario_executor.py

@@ -153,6 +153,7 @@ class ScenarioExecutor:
             verbose=verbose,
             cache_key=cache_key,
             debug=debug,
+            headless=None,
         )
         self.config = (ScenarioConfig.default_config or ScenarioConfig()).merge(config)
 
@@ -198,7 +199,7 @@ class ScenarioExecutor:
         self._state = ScenarioState(
             description=self.description,
             messages=[],
-            thread_id=str(PKSUID("thread")),
+            thread_id=str(PKSUID("scenariothread")),
             current_turn=0,
             config=self.config,
             _executor=self,

scenario/user_simulator_agent.py

@@ -37,6 +37,7 @@ class UserSimulatorAgent(AgentAdapter):
     Attributes:
         role: Always AgentRole.USER for user simulator agents
         model: LLM model identifier to use for generating user messages
+        api_base: Optional base URL where the model is hosted
         api_key: Optional API key for the model provider
         temperature: Sampling temperature for response generation
         max_tokens: Maximum tokens to generate in user messages
@@ -76,9 +77,11 @@ class UserSimulatorAgent(AgentAdapter):
         - Messages are generated in a casual, human-like style (lowercase, brief, etc.)
         - The simulator will not act as an assistant - it only generates user inputs
     """
+
     role = AgentRole.USER
 
     model: str
+    api_base: Optional[str]
     api_key: Optional[str]
     temperature: float
     max_tokens: Optional[int]
@@ -88,6 +91,7 @@ class UserSimulatorAgent(AgentAdapter):
         self,
         *,
         model: Optional[str] = None,
+        api_base: Optional[str] = None,
         api_key: Optional[str] = None,
         temperature: float = 0.0,
         max_tokens: Optional[int] = None,
@@ -99,6 +103,8 @@ class UserSimulatorAgent(AgentAdapter):
         Args:
             model: LLM model identifier (e.g., "openai/gpt-4.1").
                    If not provided, uses the default model from global configuration.
+            api_base: Optional base URL where the model is hosted. If not provided,
+                      uses the base URL from global configuration.
             api_key: API key for the model provider. If not provided,
                      uses the key from global configuration or environment.
             temperature: Sampling temperature for message generation (0.0-1.0).
@@ -128,6 +134,7 @@ class UserSimulatorAgent(AgentAdapter):
             ```
         """
         # Override the default system prompt for the user simulator agent
+        self.api_base = api_base
         self.api_key = api_key
         self.temperature = temperature
         self.max_tokens = max_tokens
@@ -144,6 +151,9 @@ class UserSimulatorAgent(AgentAdapter):
             ScenarioConfig.default_config.default_model, ModelConfig
         ):
             self.model = model or ScenarioConfig.default_config.default_model.model
+            self.api_base = (
+                api_base or ScenarioConfig.default_config.default_model.api_base
+            )
             self.api_key = (
                 api_key or ScenarioConfig.default_config.default_model.api_key
             )
@@ -222,6 +232,8 @@ Your goal (assistant) is to interact with the Agent Under Test (user) as if you
                 model=self.model,
                 messages=messages,
                 temperature=self.temperature,
+                api_key=self.api_key,
+                api_base=self.api_base,
                 max_tokens=self.max_tokens,
                 tools=[],
             ),