> ## Documentation Index
> Fetch the complete documentation index at: https://agno-v2-codex-docs-audit-20260719-0149.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Guideline-Based Agent-as-Judge Evaluation

> Steer numeric AgentAsJudgeEval scoring with additional_guidelines that require units, variant context, and technical completeness.

Demonstrates agent-as-judge scoring with additional guidelines.

```python agent_as_judge_with_guidelines.py theme={null}
"""
Guideline-Based Agent-as-Judge Evaluation
=========================================

Demonstrates agent-as-judge scoring with additional guidelines.
"""

from typing import Optional

from agno.agent import Agent
from agno.db.sqlite import SqliteDb
from agno.eval.agent_as_judge import AgentAsJudgeEval, AgentAsJudgeResult
from agno.models.openai import OpenAIChat

# ---------------------------------------------------------------------------
# Create Database
# ---------------------------------------------------------------------------
db = SqliteDb(db_file="tmp/agent_as_judge_guidelines.db")

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------
agent = Agent(
    model=OpenAIChat(id="gpt-4o"),
    instructions="You are a Tesla Model 3 product specialist. Provide detailed and helpful specifications.",
    db=db,
)

# ---------------------------------------------------------------------------
# Create Evaluation
# ---------------------------------------------------------------------------
evaluation = AgentAsJudgeEval(
    name="Product Info Quality",
    model=OpenAIChat(id="gpt-5.2"),
    criteria="Response should be informative, well-formatted, and accurate for product specifications",
    scoring_strategy="numeric",
    threshold=8,
    additional_guidelines=[
        "Must include specific numbers with proper units (mph, km/h, etc.)",
        "Should provide context for different model variants if applicable",
        "Information should be technically accurate and complete",
    ],
    db=db,
)

# ---------------------------------------------------------------------------
# Run Evaluation
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    response = agent.run("What is the maximum speed of the Tesla Model 3?")
    result: Optional[AgentAsJudgeResult] = evaluation.run(
        input="What is the maximum speed?",
        output=str(response.content),
        print_results=True,
    )
    assert result is not None, "Evaluation should return a result"

    print("Database Results:")
    eval_runs = db.get_eval_runs()
    print(f"Total evaluations stored: {len(eval_runs)}")
    if eval_runs:
        latest = eval_runs[-1]
        print(f"Eval ID: {latest.run_id}")
        print(f"Additional guidelines used: {len(evaluation.additional_guidelines)}")
```

## Run the Example

<Steps>
  <Snippet file="create-venv-step.mdx" />

  <Step title="Install dependencies">
    ```bash theme={null}
    uv pip install -U agno openai sqlalchemy
    ```
  </Step>

  <Step title="Export your OpenAI API key">
    <CodeGroup>
      ```bash Mac/Linux theme={null}
      export OPENAI_API_KEY="your_openai_api_key_here"
      ```

      ```bash Windows theme={null}
      $Env:OPENAI_API_KEY="your_openai_api_key_here"
      ```
    </CodeGroup>
  </Step>

  <Step title="Run the example">
    Save the code above as `agent_as_judge_with_guidelines.py`, then run:

    ```bash theme={null}
    python agent_as_judge_with_guidelines.py
    ```
  </Step>
</Steps>

Full source: [cookbook/09\_evals/agent\_as\_judge/agent\_as\_judge\_with\_guidelines.py](https://github.com/agno-agi/agno/blob/v2.7.4/cookbook/09_evals/agent_as_judge/agent_as_judge_with_guidelines.py)