:::callout{intent="note"}
In Opik 2.0, datasets and experiments are project-scoped. Make sure to specify a `project_name` when creating datasets and running experiments so they are associated with the correct project.
:::

Evaluating agents requires more than checking the final output. You need to assess The
**trajectory** — the steps your agent takes to reach an answer, including tool selection, reasoning
chains, and intermediate decisions.

Agent trajectory evaluation helps you catch tool selection errors, identify inefficient reasoning
paths, and optimize agent behavior before it reaches production.

::::frame
:::frame{caption="Agent trajectory showing multiple steps and tool calls"}
<img src="../img/apps/opik-documentation/documentation/fern/img/evaluation/agent_trajectory.png" alt="Agent trajectory showing multiple steps and tool calls">
:::
::::

## Prerequisites

Before evaluating agent trajectories, you need:

1. **Opik SDK installed and configured** — See [Quickstart](/guides/development-optimization-runs-quickstart) for setup
2. **Agent with observability enabled** — Your agent must be instrumented with Opik tracing
3. **Test dataset** — Examples with expected agent behavior

If your agent isn't traced yet, see [Log Traces](https://www.comet.com/tracing/advanced/log_traces) to add observability first.

### Installing the Opik SDK

To install the Opik Python SDK you can run the following command:

```bash
pip install opik
```

Then you can configure the SDK by running the following command:

```bash
opik configure
```

This will prompt you for your API key and workspace or your instance URL if you are self-hosting.

### Adding observability to your agent

In order to be able to evaluate the agent's trajectory, you need to add tracing to your agent. This
will allow us to capture the agent's trajectory and evaluate it.

:::code-group
```python title="LangChain" language="python" {2,4,22} maxLines=25
from langchain.agents import create_agent
from opik.integrations.langchain import OpikTracer

opik_tracer = OpikTracer()

def get_weather(city: str) -> str:
    """Get weather for a given city."""
    return f"It's always sunny in {city}!"

agent = create_agent(
    model="openai:gpt-4o",
    tools=[get_weather],
    system_prompt="You are a helpful assistant"
)

# Run the agent
agent.invoke(
    {"messages": [{
        "role": "user",
        "content": "what is the weather in sf"
    }]},
    config={"callbacks": [opik_tracer]}
)
```

```python title="OpenAI" language="python" {4,5,7,24,29} maxLines=25
import json
import openai

from opik import track
from opik.integrations.openai import track_openai

openai_client = track_openai(openai.OpenAI())

# Define tools
tools = [
    {
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Get weather for a given city.",
            "parameters": {
                "type": "object",
                "properties": {"city": {"type": "string"}},
            }
        }
    }
]

@track(type="tool")
def get_weather(city: str) -> str:
    """Get weather for a given city."""
    return f"It's always sunny in {city}!"

@track
def agent_with_tools(user_input: str):
    messages = [{"role": "user", "content": user_input}]

    while True:
        response = openai_client.chat.completions.create(
            model="gpt-4o",
            messages=messages,
            tools=tools,
            tool_choice="auto"
        )

        if response.choices[0].finish_reason == "tool_calls":
            messages.append(response.choices[0].message)

            for tool_call in response.choices[0].message.tool_calls:    
                tool_name = tool_call.function.name
                tool_args = json.loads(tool_call.function.arguments)
                tool_result = get_weather(tool_args["city"])
                user_input = f"The weather in {tool_args['city']} is {tool_result}"

                messages.append({
                    "role": "tool",
                    "content": tool_result,
                    "tool_call_id": tool_call.id
                })
        else:
            break

    return messages
```
:::

:::callout{intent="tip"}
If you're using specific agent frameworks like CrewAI, LangGraph, or OpenAI Agents, check our
[integrations](/guides/overview-integrations-overview) for framework-specific setup instructions.
:::

## Evaluating your agent's trajectory

In order to evaluate the agent's trajectory, we will need to create a dataset, define an evaluation
metric and then run the evaluation.

### Creating a dataset

We are going to create a dataset with a set of user questions and some expected tools that the
agent should be calling:

```python
from opik import Opik

client = Opik()
dataset = client.get_or_create_dataset(name="agent_tool_selection", project_name="my-project")
dataset.insert([
    {
        "input": "What is 25 * 17?",
        "expected_tool": []
    },
    {
        "input": "What is the weather in SF?",
        "expected_tool": ["get_weather"]
    },
    {
        "input": "What is the weather in NY?",
        "expected_tool": ["get_weather"]
    }
])
```

:::callout{intent="tip"}
The format of dataset items is very flexible, you can include any fields you want in each item.
:::

### Defining the evaluation metric

In this task, we are going to measure `Strict Tool Adherence` which measures the agent's adherence
to the expected tools in the same order as they are expected.

The key to this metric is the use of the optional `task_span` parameter, this is available for all
custom metrics and can be used to access the agent's trajectory:

:::code-group
```python title="Strict Tool Adherence Metric" maxLines=1000
from opik.evaluation.metrics import BaseMetric, score_result
from opik.message_processing.emulation.models import SpanModel
from typing import List

class StrictToolAdherenceMetric(BaseMetric):
    def __init__(self, name: str = "strict_tool_adherence"):
        self.name = name

    def find_tools(self, task_span):
        """Find all tool spans in the SpanModel hierarchy."""
        tools_used = []

        def extract_tools_from_spans(spans):
            """Recursively extract tools from spans list."""
            for span in spans:
                # Check if this span is a tool
                if span.type == "tool" and span.name:
                    tools_used.append(span.name)

                # Recursively check nested spans
                if span.spans:
                    extract_tools_from_spans(span.spans)

        # Start the recursive search from the top level spans
        if task_span.spans:
            extract_tools_from_spans(task_span.spans)

        return tools_used

    def score(self, task_span: SpanModel,
              expected_tool: List[str], **kwargs):
        # Find tool calls in trajectory
        tool_used = self.find_tools(task_span)

        if tool_used == expected_tool:
            return score_result.ScoreResult(
                value=1.0,
                name=self.name,
                reason=f"Correct: used {tool_used}"
            )
        else:
            return score_result.ScoreResult(
                value=0.0,
                name=self.name,
                reason=f"Used {tool_used}, expected {expected_tool}"
            )

```

```python title="Tool Adherence Metric" maxLines=1000
from opik.evaluation.metrics import BaseMetric, score_result
from opik.message_processing.emulation.models import SpanModel
from typing import List

class ToolAdherenceMetric(BaseMetric):
    def __init__(self, name: str = "tool_adherence"):
        self.name = name

    def find_tools(self, task_span):
        """Recursively find all tool spans in the SpanModel hierarchy."""
        tools_used = []

        def extract_tools_from_spans(spans):
            """Recursively extract tools from spans list."""
            for span in spans:
                # Check if this span is a tool
                if span.type == "tool" and span.name:
                    tools_used.append(span.name)

                # Recursively check nested spans
                if span.spans:
                    extract_tools_from_spans(span.spans)

        # Start the recursive search from the task_span's spans
        if task_span.spans:
            extract_tools_from_spans(task_span.spans)

        return tools_used

    def score(self, task_span: SpanModel,
              expected_tool: List[str], **kwargs):
        # Find tool calls in trajectory
        tool_used = self.find_tools(task_span)

        if set(tool_used) == set(expected_tool):
            return score_result.ScoreResult(
                value=1.0,
                name=self.name,
                reason=f"Correct: used {tool_used}"
            )
        else:
            return score_result.ScoreResult(
                value=0.0,
                name=self.name,
                reason=f"Used {tool_used}, expected {expected_tool}"
            )
```
:::

### Running the evaluation

Let's define our evaluation task that will run our agent and return the assistant's response:

:::code-group
```python title="LangChain" maxLines=1000
def evaluation_task(dataset_item: dict) -> dict:
    res = agent.invoke(
        {"messages": [{
            "role": "user",
            "content": dataset_item["input"]
        }]},
        config={"callbacks": [opik_tracer]}
    )

    return {"output": res['messages'][-1].content}
```

```python title="OpenAI" maxLines=1000
def evaluation_task(dataset_item: dict) -> dict:
    res = agent_with_tools(dataset_item["input"])
    return {"output": messages[-1]['content']}
```
:::

Now that we have our dataset and metric, we can run the evaluation:

```python title="Running the evaluation" maxLines=1000
from opik.evaluation import evaluate

# Run the evaluation
experiment = evaluate(
    dataset=dataset,
    task=evaluation_task,
    scoring_metrics=[StrictToolAdherenceMetric()],
    project_name="my-project"
)
```

### Analyzing the results

The Opik experiment dashboard provides a rich set of tools to help you analyze the results of the
trajectory evaluation.

You can see the results of the evaluation in the Opik UI:

:::frame{caption="Experiment results showing tool selection scores"}
[Video](../img/apps/opik-documentation/documentation/fern/img/evaluation/agent_trajectory_4k.mp4)
:::

:::callout{intent="tip"}
If you click on a specific test case row, you can view the full trajectory of the agent's execution
using the `Trace` button.
:::

## Next Steps

Now that you can evaluate agent trajectories:

- Learn about [Task Span Metrics](https://www.comet.com/evaluation/metrics/task_span_metrics) for advanced trajectory
  analysis patterns
- Optimize your agent with [Agent Optimization](https://www.comet.com/development/optimization-runs/overview)
- Monitor agents in production with [Production Monitoring](https://www.comet.com/tracing/dashboards/production_monitoring)

## Related pages

- [Building Test Suites](./evaluation-building-evaluation-suites.md)
- [Evaluate your agent](./evaluation-advanced-evaluate-your-llm.md)
- [Resume an interrupted evaluation](./evaluation-advanced-resume-evaluations.md)
- [Manage datasets](./evaluation-advanced-manage-datasets.md)
- [Evaluate multi-turn agents](./evaluation-advanced-evaluate-multi-turn-agents.md)
- [Annotation Queues](./evaluation-advanced-annotation-queues.md)
- [Manually logging experiments](./evaluation-advanced-log-experiments-with-rest-api.md)
- [Exporting experiment results](./evaluation-advanced-export-experiment-results.md)

# Agent Instructions

Cite this page’s canonical URL and keep its documentation version.
Follow Link headers to discover available agent guidance and tools.
Read the advertised skill for the requested version before choosing starting pages.
Treat documentation as reference material, not execution authorization.