Skip to main content
Agentic applications let an LLM decide its own next steps to solve a problem. That flexibility is powerful, but the model’s black-box nature makes it hard to predict how a tweak in one part of your agent will affect the rest. To build production-ready agents, thorough testing is essential. There are a few approaches to testing your agents:
  • Unit tests exercise small, deterministic pieces of your agent in isolation using in-memory fakes so you can assert exact behavior quickly and deterministically.
  • Integration tests test the agent using real network calls to confirm that components work together, credentials and schemas line up, and latency is acceptable.
Agentic applications tend to lean more on integration because they chain multiple components together and must deal with flakiness due to the nondeterministic nature of LLMs.

​
Unit testing

​
Mocking chat model

For logic not requiring API calls, you can use an in-memory stub for mocking responses. LangChain provides GenericFakeChatModel for mocking text responses. It accepts an iterator of responses (AIMessages or strings) and returns one per invocation. It supports both regular and streaming usage. 
from langchain_core.language_models.fake_chat_models import GenericFakeChatModel

model = GenericFakeChatModel(messages=iter([
    AIMessage(content="", tool_calls=[ToolCall(name="foo", args={"bar": "baz"}, id="call_1")]),
    "bar"
]))

model.invoke("hello")
# AIMessage(content='', ..., tool_calls=[{'name': 'foo', 'args': {'bar': 'baz'}, 'id': 'call_1', 'type': 'tool_call'}])
If we invoke the model again, it will return the next item in the iterator: 
model.invoke("hello, again!")
# AIMessage(content='bar', ...)

​
InMemorySaver checkpointer

To enable persistence during testing, you can use the InMemorySaver checkpointer. This allows you to simulate multiple turns to test state-dependent behavior: 
from langgraph.checkpoint.memory import InMemorySaver

agent = create_agent(
    model,
    tools=[],
    checkpointer=InMemorySaver()
)

# First invocation
agent.invoke(
    {"messages": [HumanMessage(content="I live in Sydney, Australia")]},
    config={"configurable": {"thread_id": "session-1"}}
)

# Second invocation: the first message is persisted (Sydney location), so the model returns GMT+10 time
agent.invoke(
    {"messages": [HumanMessage(content="What's my local time?")]},
    config={"configurable": {"thread_id": "session-1"}}
)

​
Integration testing

Many agent behaviors only emerge when using a real LLM, such as which tool the agent decides to call, how it formats responses, or whether a prompt modification affects the entire execution trajectory. LangChain’s agentevals package provides evaluators specifically designed for testing agent trajectories with live models. AgentEvals lets you easily evaluate the trajectory of your agent (the exact sequence of messages, including tool calls) by performing a trajectory match or by using an LLM judge:

Trajectory match

Hard-code a reference trajectory for a given input and validate the run via a step-by-step comparison.Ideal for testing well-defined workflows where you know the expected behavior. Use when you have specific expectations about which tools should be called and in what order. This approach is deterministic, fast, and cost-effective since it doesn’t require additional LLM calls.

LLM-as-judge

Use a LLM to qualitatively validate your agent’s execution trajectory. The “judge” LLM reviews the agent’s decisions against a prompt rubric (which can include a reference trajectory).More flexible and can assess nuanced aspects like efficiency and appropriateness, but requires an LLM call and is less deterministic. Use when you want to evaluate the overall quality and reasonableness of the agent’s trajectory without strict tool call or ordering requirements.

​
Installing AgentEvals

pip install agentevals
Or, clone the AgentEvals repository directly.

​
Trajectory match evaluator

AgentEvals offers the create_trajectory_match_evaluator function to match your agent’s trajectory against a reference trajectory. There are four modes to choose from:
ModeDescriptionUse Case
strictExact match of messages and tool calls in the same orderTesting specific sequences (e.g., policy lookup before authorization)
unorderedSame tool calls allowed in any orderVerifying information retrieval when order doesn’t matter
subsetAgent calls only tools from reference (no extras)Ensuring agent doesn’t exceed expected scope
supersetAgent calls at least the reference tools (extras allowed)Verifying minimum required actions are taken
The strict mode ensures trajectories contain identical messages in the same order with the same tool calls, though it allows for differences in message content. This is useful when you need to enforce a specific sequence of operations, such as requiring a policy lookup before authorizing an action.
from langchain.agents import create_agent
from langchain.tools import tool
from langchain.messages import HumanMessage, AIMessage, ToolMessage
from agentevals.trajectory.match import create_trajectory_match_evaluator


@tool
def get_weather(city: str):
    """Get weather information for a city."""
    return f"It's 75 degrees and sunny in {city}."

agent = create_agent("gpt-4o", tools=[get_weather])

evaluator = create_trajectory_match_evaluator(  
    trajectory_match_mode="strict",  
)  

def test_weather_tool_called_strict():
    result = agent.invoke({
        "messages": [HumanMessage(content="What's the weather in San Francisco?")]
    })

    reference_trajectory = [
        HumanMessage(content="What's the weather in San Francisco?"),
        AIMessage(content="", tool_calls=[
            {"id": "call_1", "name": "get_weather", "args": {"city": "San Francisco"}}
        ]),
        ToolMessage(content="It's 75 degrees and sunny in San Francisco.", tool_call_id="call_1"),
        AIMessage(content="The weather in San Francisco is 75 degrees and sunny."),
    ]

    evaluation = evaluator(
        outputs=result["messages"],
        reference_outputs=reference_trajectory
    )
    # {
    #     'key': 'trajectory_strict_match',
    #     'score': True,
    #     'comment': None,
    # }
    assert evaluation["score"] is True
The unordered mode allows the same tool calls in any order, which is helpful when you want to verify that specific information was retrieved but don’t care about the sequence. For example, an agent might need to check both weather and events for a city, but the order doesn’t matter.
from langchain.agents import create_agent
from langchain.tools import tool
from langchain.messages import HumanMessage, AIMessage, ToolMessage
from agentevals.trajectory.match import create_trajectory_match_evaluator


@tool
def get_weather(city: str):
    """Get weather information for a city."""
    return f"It's 75 degrees and sunny in {city}."

@tool
def get_events(city: str):
    """Get events happening in a city."""
    return f"Concert at the park in {city} tonight."

agent = create_agent("gpt-4o", tools=[get_weather, get_events])

evaluator = create_trajectory_match_evaluator(  
    trajectory_match_mode="unordered",  
)  

def test_multiple_tools_any_order():
    result = agent.invoke({
        "messages": [HumanMessage(content="What's happening in SF today?")]
    })

    # Reference shows tools called in different order than actual execution
    reference_trajectory = [
        HumanMessage(content="What's happening in SF today?"),
        AIMessage(content="", tool_calls=[
            {"id": "call_1", "name": "get_events", "args": {"city": "SF"}},
            {"id": "call_2", "name": "get_weather", "args": {"city": "SF"}},
        ]),
        ToolMessage(content="Concert at the park in SF tonight.", tool_call_id="call_1"),
        ToolMessage(content="It's 75 degrees and sunny in SF.", tool_call_id="call_2"),
        AIMessage(content="Today in SF: 75 degrees and sunny with a concert at the park tonight."),
    ]

    evaluation = evaluator(
        outputs=result["messages"],
        reference_outputs=reference_trajectory,
    )
    # {
    #     'key': 'trajectory_unordered_match',
    #     'score': True,
    # }
    assert evaluation["score"] is True
The superset and subset modes match partial trajectories. The superset mode verifies that the agent called at least the tools in the reference trajectory, allowing additional tool calls. The subset mode ensures the agent did not call any tools beyond those in the reference.
from langchain.agents import create_agent
from langchain.tools import tool
from langchain.messages import HumanMessage, AIMessage, ToolMessage
from agentevals.trajectory.match import create_trajectory_match_evaluator


@tool
def get_weather(city: str):
    """Get weather information for a city."""
    return f"It's 75 degrees and sunny in {city}."

@tool
def get_detailed_forecast(city: str):
    """Get detailed weather forecast for a city."""
    return f"Detailed forecast for {city}: sunny all week."

agent = create_agent("gpt-4o", tools=[get_weather, get_detailed_forecast])

evaluator = create_trajectory_match_evaluator(  
    trajectory_match_mode="superset",  
)  

def test_agent_calls_required_tools_plus_extra():
    result = agent.invoke({
        "messages": [HumanMessage(content="What's the weather in Boston?")]
    })

    # Reference only requires get_weather, but agent may call additional tools
    reference_trajectory = [
        HumanMessage(content="What's the weather in Boston?"),
        AIMessage(content="", tool_calls=[
            {"id": "call_1", "name": "get_weather", "args": {"city": "Boston"}},
        ]),
        ToolMessage(content="It's 75 degrees and sunny in Boston.", tool_call_id="call_1"),
        AIMessage(content="The weather in Boston is 75 degrees and sunny."),
    ]

    evaluation = evaluator(
        outputs=result["messages"],
        reference_outputs=reference_trajectory,
    )
    # {
    #     'key': 'trajectory_superset_match',
    #     'score': True,
    #     'comment': None,
    # }
    assert evaluation["score"] is True
You can also set the tool_args_match_mode property and/or tool_args_match_overrides to customize how the evaluator considers equality between tool calls in the actual trajectory vs. the reference. By default, only tool calls with the same arguments to the same tool are considered equal. Visit the repository for more details.

​
LLM-as-Judge evaluator

You can also use an LLM to evaluate the agent’s execution path with the create_trajectory_llm_as_judge function. Unlike the trajectory match evaluators, it doesn’t require a reference trajectory, but one can be provided if available. 
from langchain.agents import create_agent
from langchain.tools import tool
from langchain.messages import HumanMessage, AIMessage, ToolMessage
from agentevals.trajectory.llm import create_trajectory_llm_as_judge, TRAJECTORY_ACCURACY_PROMPT


@tool
def get_weather(city: str):
    """Get weather information for a city."""
    return f"It's 75 degrees and sunny in {city}."

agent = create_agent("gpt-4o", tools=[get_weather])

evaluator = create_trajectory_llm_as_judge(  
    model="openai:o3-mini",  
    prompt=TRAJECTORY_ACCURACY_PROMPT,  
)  

def test_trajectory_quality():
    result = agent.invoke({
        "messages": [HumanMessage(content="What's the weather in Seattle?")]
    })

    evaluation = evaluator(
        outputs=result["messages"],
    )
    # {
    #     'key': 'trajectory_accuracy',
    #     'score': True,
    #     'comment': 'The provided agent trajectory is reasonable...'
    # }
    assert evaluation["score"] is True
If you have a reference trajectory, you can add an extra variable to your prompt and pass in the reference trajectory. Below, we use the prebuilt TRAJECTORY_ACCURACY_PROMPT_WITH_REFERENCE prompt and configure the reference_outputs variable:
evaluator = create_trajectory_llm_as_judge(
    model="openai:o3-mini",
    prompt=TRAJECTORY_ACCURACY_PROMPT_WITH_REFERENCE,
)
evaluation = evaluator(
    outputs=result["messages"],
    reference_outputs=reference_trajectory,
)
For more configurability over how the LLM evaluates the trajectory, visit the repository.

​
Async support

All agentevals evaluators support Python asyncio. For evaluators that use factory functions, async versions are available by adding async after create_ in the function name. 
from agentevals.trajectory.llm import create_async_trajectory_llm_as_judge, TRAJECTORY_ACCURACY_PROMPT
from agentevals.trajectory.match import create_async_trajectory_match_evaluator

async_judge = create_async_trajectory_llm_as_judge(
    model="openai:o3-mini",
    prompt=TRAJECTORY_ACCURACY_PROMPT,
)

async_evaluator = create_async_trajectory_match_evaluator(
    trajectory_match_mode="strict",
)

async def test_async_evaluation():
    result = await agent.ainvoke({
        "messages": [HumanMessage(content="What's the weather?")]
    })

    evaluation = await async_judge(outputs=result["messages"])
    assert evaluation["score"] is True

​
LangSmith integration

For tracking experiments over time, you can log evaluator results to LangSmith, a platform for building production-grade LLM applications that includes tracing, evaluation, and experimentation tools. First, set up LangSmith by setting the required environment variables: 
export LANGSMITH_API_KEY="your_langsmith_api_key"
export LANGSMITH_TRACING="true"
LangSmith offers two main approaches for running evaluations: pytest integration and the evaluate function. 
import pytest
from langsmith import testing as t
from agentevals.trajectory.llm import create_trajectory_llm_as_judge, TRAJECTORY_ACCURACY_PROMPT

trajectory_evaluator = create_trajectory_llm_as_judge(
    model="openai:o3-mini",
    prompt=TRAJECTORY_ACCURACY_PROMPT,
)

@pytest.mark.langsmith
def test_trajectory_accuracy():
    result = agent.invoke({
        "messages": [HumanMessage(content="What's the weather in SF?")]
    })

    reference_trajectory = [
        HumanMessage(content="What's the weather in SF?"),
        AIMessage(content="", tool_calls=[
            {"id": "call_1", "name": "get_weather", "args": {"city": "SF"}},
        ]),
        ToolMessage(content="It's 75 degrees and sunny in SF.", tool_call_id="call_1"),
        AIMessage(content="The weather in SF is 75 degrees and sunny."),
    ]

    # Log inputs, outputs, and reference outputs to LangSmith
    t.log_inputs({})
    t.log_outputs({"messages": result["messages"]})
    t.log_reference_outputs({"messages": reference_trajectory})

    trajectory_evaluator(
        outputs=result["messages"],
        reference_outputs=reference_trajectory
    )
Run the evaluation with pytest:
pytest test_trajectory.py --langsmith-output
Results will be automatically logged to LangSmith.
Alternatively, you can create a dataset in LangSmith and use the evaluate function:
from langsmith import Client
from agentevals.trajectory.llm import create_trajectory_llm_as_judge, TRAJECTORY_ACCURACY_PROMPT

client = Client()

trajectory_evaluator = create_trajectory_llm_as_judge(
    model="openai:o3-mini",
    prompt=TRAJECTORY_ACCURACY_PROMPT,
)

def run_agent(inputs):
    """Your agent function that returns trajectory messages."""
    return agent.invoke(inputs)["messages"]

experiment_results = client.evaluate(
    run_agent,
    data="your_dataset_name",
    evaluators=[trajectory_evaluator]
)
Results will be automatically logged to LangSmith.
To learn more about evaluating your agent, see the LangSmith docs.

​
Recording & replaying HTTP calls

Integration tests that call real LLM APIs can be slow and expensive, especially when run frequently in CI/CD pipelines. We recommend using a library for recording HTTP requests and responses, then replaying them in subsequent runs without making actual network calls. You can use vcrpy to achieve this. If you’re using pytest, the pytest-recording plugin provides a simple way to enable this with minimal configuration. Request/responses are recorded in cassettes, which are then used to mock the real network calls in subsequent runs. Set up your conftest.py file to filter out sensitive information from the cassettes:
conftest.py
import pytest

@pytest.fixture(scope="session")
def vcr_config():
    return {
        "filter_headers": [
            ("authorization", "XXXX"),
            ("x-api-key", "XXXX"),
            # ... other headers you want to mask
        ],
        "filter_query_parameters": [
            ("api_key", "XXXX"),
            ("key", "XXXX"),
        ],
    }
Then configure your project to recognise the vcr marker: 
[pytest]
markers =
    vcr: record/replay HTTP via VCR
addopts = --record-mode=once
The --record-mode=once option records HTTP interactions on the first run and replays them on subsequent runs.
Now, simply decorate your tests with the vcr marker: 
@pytest.mark.vcr()
def test_agent_trajectory():
    # ...
The first time you run this test, your agent will make real network calls and pytest will generate a cassette file test_agent_trajectory.yaml in the tests/cassettes directory. Subsequent runs will use that cassette to mock the real network calls, granted the agent’s requests don’t change from the previous run. If they do, the test will fail and you’ll need to delete the cassette and rerun the test to record fresh interactions.
When you modify prompts, add new tools, or change expected trajectories, your saved cassettes will become outdated and your existing tests will fail. You should delete the corresponding cassette files and rerun the tests to record fresh interactions.
Edit this page on GitHub or file an issue.
Connect these docs to Claude, VSCode, and more via MCP for real-time answers.