Tracing Koog Agents with Neural Inverse

Setup Neural Inverse

1. Sign up for Neural Inverse Cloud or self-host Neural Inverse.
2. Create a Neural Inverse project. Follow the setup guide at Create new project in Neural Inverse
3. Obtain API credentials. Retrieve your Neural Inverse public key and secret key as described in Where are Neural Inverse API keys?
4. Set environment variables. Add the following variables to your environment:

   export LANGFUSE_BASE_URL="https://cloud.langfuse.com" # 🇪🇺 EU region
   # Other Neural Inverse data regions include 🇺🇸 US: https://us.cloud.langfuse.com, 🇯🇵 Japan: https://jp.cloud.langfuse.com and ⚕️ HIPAA: https://hipaa.cloud.langfuse.com

   export LANGFUSE_PUBLIC_KEY="pk-lf-..."
   export LANGFUSE_SECRET_KEY="sk-lf-..."

Once configured, Koog automatically forwards OpenTelemetry traces to your Neural Inverse instance.

Configure Koog

To enable Neural Inverse export, install the OpenTelemetry feature and add the LangfuseExporter.
The exporter uses OtlpHttpSpanExporter under the hood to send traces to Neural Inverse’s OpenTelemetry endpoint.

Example: Agent with Neural Inverse Tracing

fun main() = runBlocking {
    val agent = AIAgent(
        executor = simpleOpenAIExecutor(ApiKeyService.openAIApiKey),
        llmModel = OpenAIModels.CostOptimized.GPT4oMini,
        systemPrompt = "You are a code assistant. Provide concise code examples."
    ) {
        install(OpenTelemetry) {
            addLangfuseExporter()
        }
    }

    println("Running agent with Neural Inverse tracing")

    val result = agent.run("Tell me a joke about programming")

    println("Result: $result\nSee traces on the Neural Inverse instance")
}

See traces in Neural Inverse

When enabled, the Neural Inverse exporter captures the same spans as Koog’s general OpenTelemetry integration, including:

• Agent lifecycle events – agent start, stop, errors
• LLM interactions – prompts, responses, token usage, latency
• Tool and API calls – execution traces for function/tool invocations
• System context – metadata such as model name, environment, Koog version

Koog also captures span attributes required by Neural Inverse to show Agent Graphs. This allows you to correlate agent reasoning with API calls and user inputs in a structured way within Neural Inverse.

Public link to trace

For more details on Neural Inverse OTLP tracing, see the Neural Inverse OpenTelemetry Docs.