online-evals

Config Online Evaluations

Attach judges to config variations for automatic quality scoring using LLM-as-a-judge methodology. Judges evaluate responses and return scores between 0.0 and 1.0.

Prerequisites

• LaunchDarkly account with AgentControl enabled
• API access token with write permissions
• Existing config with variations (use configs-create skill)
• For automatic metric recording and the consolidated judge-result API: Python AI SDK v0.20.0+ or Node.js AI SDK v0.20.0+

API Key Detection

1. Check environment variables - LAUNCHDARKLY_API_KEY, LAUNCHDARKLY_API_TOKEN, LD_API_KEY
2. Check MCP config - Claude: ~/.claude/config.json -> mcpServers.launchdarkly.env.LAUNCHDARKLY_API_KEY
3. Prompt user - Only if detection fails

Core Concepts

What Are Judges?

Judges are specialized configs in judge mode that evaluate responses from other configs. They use an LLM to score outputs and return structured results:

{
  "score": 0.85,
  "reasoning": "Answered correctly with one minor omission"
}

Built-in Judges

LaunchDarkly provides three pre-configured judges:

Completion Mode Only

Judges can only be attached to completion mode configs in the UI. For agent mode or custom pipelines, use programmatic evaluation via the SDK.

Restrictions

• Cannot attach judges to judges (no recursion)
• Cannot attach multiple judges with the same metric key to a single variation
• Cannot view/edit model parameters or tools on judge variations

Workflow

Step 1: Create Custom Judges (Optional)

For domain-specific evaluation, create judge configs:

# Create judge config
curl -X POST "https://app.launchdarkly.com/api/v2/projects/{projectKey}/ai-configs" \
  -H "Authorization: {api_token}" \
  -H "Content-Type: application/json" \
  -H "LD-API-Version: beta" \
  -d '{
    "key": "security-judge",
    "name": "Security Judge",
    "mode": "judge",
    "evaluationMetricKey": "security",
    "isInverted": false
  }'

Note: Set isInverted: true for metrics like toxicity where 0.0 is better.

Then add a variation with the evaluation prompt:

curl -X POST "https://app.launchdarkly.com/api/v2/projects/{projectKey}/ai-configs/security-judge/variations" \
  -H "Authorization: {api_token}" \
  -H "Content-Type: application/json" \
  -H "LD-API-Version: beta" \
  -d '{
    "key": "default",
    "name": "Default",
    "messages": [
      {
        "role": "system",
        "content": "You are a security auditor. Score from 0.0 to 1.0:\n- 1.0: No security issues\n- 0.7-0.9: Minor issues\n- 0.4-0.6: Moderate issues\n- 0.1-0.3: Serious vulnerabilities\n- 0.0: Critical vulnerabilities\n\nCheck for: SQL injection, XSS, hardcoded secrets, command injection."
      }
    ],
    "modelConfigKey": "OpenAI.gpt-4o-mini",
    "model": {
      "parameters": {
        "temperature": 0.3
      }
    }
  }'

Step 2: Attach Judges to Variations

Use the variation PATCH endpoint:

curl -X PATCH "https://app.launchdarkly.com/api/v2/projects/{projectKey}/ai-configs/{configKey}/variations/{variationKey}" \
  -H "Authorization: {api_token}" \
  -H "Content-Type: application/json" \
  -H "LD-API-Version: beta" \
  -d '{
    "judgeConfiguration": {
      "judges": [
        {"judgeConfigKey": "security-judge", "samplingRate": 1.0},
        {"judgeConfigKey": "api-contract-judge", "samplingRate": 0.5}
      ]
    }
  }'

Important: The judges array replaces all existing judge attachments. An empty array removes all judges.

Step 3: Set Fallthrough on Judges

Each judge config needs its fallthrough set to the enabled variation. Configs default to the "disabled" variation (index 0).

Note: turnTargetingOn does not work for configs. Use updateFallthroughVariationOrRollout instead.

# First get the variation ID for "Default" from GET targeting response
curl -X PATCH "https://app.launchdarkly.com/api/v2/projects/{projectKey}/ai-configs/security-judge/targeting" \
  -H "Authorization: {api_token}" \
  -H "Content-Type: application/json; domain-model=launchdarkly.semanticpatch" \
  -H "LD-API-Version: beta" \
  -d '{
    "environmentKey": "production",
    "instructions": [{
      "kind": "updateFallthroughVariationOrRollout",
      "variationId": "your-default-variation-uuid"
    }]
  }'

Python Implementation

import requests
import os
from typing import Optional

class AIConfigJudges:
    """Manager for config judge attachments"""

    def __init__(self, api_token: str, project_key: str):
        self.api_token = api_token
        self.project_key = project_key
        self.base_url = "https://app.launchdarkly.com/api/v2"
        self.headers = {
            "Authorization": api_token,
            "Content-Type": "application/json",
            "LD-API-Version": "beta"
        }

    def attach_judges(self, config_key: str, variation_key: str,
                      judges: list[dict]) -> dict:
        """
        Attach judges to a variation.

        Args:
            config_key: config key
            variation_key: Variation key
            judges: List of {"judgeConfigKey": str, "samplingRate": float}
        """
        url = f"{self.base_url}/projects/{self.project_key}/ai-configs/{config_key}/variations/{variation_key}"

        response = requests.patch(url, headers=self.headers, json={
            "judgeConfiguration": {"judges": judges}
        })

        if response.status_code == 200:
            print(f"[OK] Attached {len(judges)} judges to {config_key}/{variation_key}")
            return response.json()
        print(f"[ERROR] {response.status_code}: {response.text}")
        return {}

    def create_judge(self, key: str, name: str, metric_key: str,
                     system_prompt: str, model: str = "OpenAI.gpt-4o-mini",
                     is_inverted: bool = False) -> dict:
        """
        Create a judge config.

        Args:
            key: Judge config key
            name: Display name
            metric_key: Metric key for scoring (appears as $ld:ai:judge:{metric_key})
            system_prompt: Evaluation instructions
            is_inverted: True if lower scores are better (e.g., toxicity)
        """
        # Create config
        config_url = f"{self.base_url}/projects/{self.project_key}/ai-configs"
        response = requests.post(config_url, headers=self.headers, json={
            "key": key,
            "name": name,
            "mode": "judge",
            "evaluationMetricKey": metric_key,
            "isInverted": is_inverted
        })

        if response.status_code not in [200, 201]:
            print(f"[ERROR] Creating config: {response.text}")
            return {}

        # Create variation
        var_url = f"{self.base_url}/projects/{self.project_key}/ai-configs/{key}/variations"
        response = requests.post(var_url, headers=self.headers, json={
            "key": "default",
            "name": "Default",
            "messages": [{"role": "system", "content": system_prompt}],
            "modelConfigKey": model,
            "model": {"parameters": {"temperature": 0.3}}
        })

        if response.status_code in [200, 201]:
            print(f"[OK] Created judge: {key}")
            return response.json()
        print(f"[ERROR] Creating variation: {response.text}")
        return {}

    def set_fallthrough(self, config_key: str, environment: str,
                        variation_key: str = "default") -> bool:
        """
        Set fallthrough to enable a judge config.

        Note: turnTargetingOn doesn't work for configs. Instead, set the
        fallthrough from disabled (index 0) to the enabled variation.
        """
        # Get variation ID
        url = f"{self.base_url}/projects/{self.project_key}/ai-configs/{config_key}/targeting"
        response = requests.get(url, headers=self.headers)

        if response.status_code != 200:
            print(f"[ERROR] {response.status_code}: {response.text}")
            return False

        targeting = response.json()
        variation_id = None
        for var in targeting.get("variations", []):
            if var.get("key") == variation_key or var.get("name") == variation_key:
                variation_id = var.get("_id")
                break

        if not variation_id:
            print(f"[ERROR] Variation '{variation_key}' not found")
            return False

        # Set fallthrough
        response = requests.patch(url, headers={
            **self.headers,
            "Content-Type": "application/json; domain-model=launchdarkly.semanticpatch"
        }, json={
            "environmentKey": environment,
            "instructions": [{
                "kind": "updateFallthroughVariationOrRollout",
                "variationId": variation_id
            }]
        })

        if response.status_code == 200:
            print(f"[OK] Fallthrough set for {config_key}")
            return True
        print(f"[ERROR] {response.status_code}: {response.text}")
        return False

SDK: Automatic Evaluation

When using create_model() + run(), attached judges evaluate automatically:

import os
import json
import asyncio
import ldclient
from ldclient import Context
from ldclient.config import Config
from ldai import LDAIClient, AICompletionConfigDefault

sdk_key = os.getenv('LAUNCHDARKLY_SDK_KEY')
ai_config_key = os.getenv('LAUNCHDARKLY_AI_CONFIG_KEY', 'sample-ai-config')

async def async_main():
    ldclient.set_config(Config(sdk_key))
    aiclient = LDAIClient(ldclient.get())

    context = (
        Context.builder('example-user-key')
        .kind('user')
        .name('Sandy')
        .build()
    )

    default_value = AICompletionConfigDefault(enabled=False)

    # create_model() initializes with judges from Config
    model = await aiclient.create_model(ai_config_key, context, default_value, {})

    if not model:
        print(f"agent configuration not enabled for: {ai_config_key}")
        return

    user_input = 'How can LaunchDarkly help me?'

    # run() automatically evaluates with attached judges
    result = await model.run(user_input)
    print("Response:", result.content)

    # Await evaluation results
    if result.evaluations and len(result.evaluations) > 0:
        eval_results = await asyncio.gather(*result.evaluations)
        results_to_display = [
            r.to_dict() if r is not None else "not evaluated"
            for r in eval_results
        ]
        print("Judge results:")
        print(json.dumps(results_to_display, indent=2, default=str))

    # Always flush events before closing — trailing events are at risk of being
    # lost otherwise, in short-lived scripts and long-running services alike.
    ldclient.get().flush()
    ldclient.get().close()

SDK: Direct Judge Evaluation

For agent mode or custom pipelines, evaluate input/output pairs directly:

import os
import json
import asyncio
import ldclient
from ldclient import Context
from ldclient.config import Config
from ldai import LDAIClient, AIJudgeConfigDefault

sdk_key = os.getenv('LAUNCHDARKLY_SDK_KEY')
judge_key = os.getenv('LAUNCHDARKLY_AI_JUDGE_KEY', 'sample-ai-judge-accuracy')

async def async_main():
    ldclient.set_config(Config(sdk_key))
    aiclient = LDAIClient(ldclient.get())

    context = (
        Context.builder('example-user-key')
        .kind('user')
        .name('Sandy')
        .build()
    )

    judge_default_value = AIJudgeConfigDefault(enabled=False)

    # Get judge configuration from LaunchDarkly
    judge = aiclient.create_judge(judge_key, context, judge_default_value)

    if not judge:
        print(f"agent judge configuration not enabled for key: {judge_key}")
        return

    input_text = 'You are a helpful assistant. How can you help me?'
    output_text = 'I can answer any question you have.'

    # Evaluate the input/output pair — returns a JudgeResult.
    judge_result = await judge.evaluate(input_text, output_text)

    if not judge_result.sampled:
        print("Judge evaluation was skipped (sample rate or configuration issue)")
        return

    # Track the consolidated result on the Config tracker if needed:
    # tracker = ai_config.create_tracker()
    # tracker.track_judge_result(judge_result)

    print("Judge Result:")
    print(json.dumps(judge_result.to_dict(), default=str))

    # Always flush events before closing — trailing events are at risk of being
    # lost otherwise, in short-lived scripts and long-running services alike.
    ldclient.get().flush()
    ldclient.get().close()

Note: Direct evaluation does not automatically record metrics. Obtain a tracker via ai_config.create_tracker() / aiConfig.createTracker() and call tracker.track_judge_result(result) / tracker.trackJudgeResult(result) to record scores for the config you're evaluating.

Sampling Rates

Each evaluated response sends an additional request to your model provider, increasing token usage and costs. Start with a lower sampling percentage and increase only if you need more evaluation coverage.

You can adjust sampling rates at any time from the Judges section of a variation, or disable a judge by setting its sampling to 0%.

Viewing Results

1. Navigate to configs > select your config
2. Click Monitoring tab
3. Select Evaluator metrics from dropdown
4. View scores by variation and time range

Results appear within 1-2 minutes of evaluation.

Use in Guardrails and Experiments

Evaluation metrics integrate with:

• Guarded rollouts: Pause/revert when scores fall below threshold
• Experiments: Compare variations using evaluation metrics as goals

Error Handling

Next Steps

After attaching judges:

1. Set fallthrough on judge configs to an enabled variation (required)
2. Monitor results in Monitoring tab
3. Adjust sampling based on cost/coverage needs
4. Set up guarded rollouts for automatic regression detection

Related Skills

• configs-create - Create configs and judges
• configs-targeting - Configure targeting rules
• configs-variations - Manage variations

References

• Online Evaluations
• Custom Judges

Python SDK examples:

• create_judge_example.py - Evaluate input/output pairs directly via create_judge + evaluate
• create_model_example.py - Automatic evaluation with create_model + run (attached judges fire during the run)

Node.js SDK examples:

• features/create-judge - Evaluate input/output pairs directly via createJudge + evaluate
• features/create-model - Automatic evaluation with createModel + run (attached judges fire during the run)