Integrate your App

Now it is time to integrate your own application to send the LLM traces over to Kubit and start getting insights from Agent Analytics.

If you haven't done so yet, please

Install Kubit Plugin to Claude Code or Cursor
(Optional) Sign Up manually, or wait for OAuth to take you through the flow.
(Optional) Check out the Example workspace and Demo Store

Overview

kubit-integrate will config your project to send LLM traces to Kubit through OpenTelemetry instrumentation:

Detect your tracing setup in the current repo, including AI Framework (e.g. Langchain, Vercel AI) and Observability SDK (Langfuse, Braintrust, Arize or Langsmith).
Ensure a Kubit session (delegating to /kubit-connect if you don't have one).
Pick a workspace — use your current one, switch to another in your org, or create a new workspace.
New workspace created using kubit-integrate skill will use a default Workspace Context derived from your email domain. To fine tune the Workspace Context to make enrichment more accurate, see Create Workspace
Mint an ingestion API key scoped to that workspace.
Write the key into your repo's .env.local or .env (gitignore-checked).
Install the Kubit SDK via your project's package manager.
Wire Kubit's span processor into your existing tracing setup — merging into your existing wiring file when one exists, or scaffolding a standalone bootstrap file when none does.

Your existing tracing setup won't be impacted since Kubit's span processor will work in parallel through OTel fan-out.

After it finishes, you run a single verification command (printed at the end) and traces should appear in Kubit once you app starts running.

Data Flow

LLM traces generated through instrumentation done by your existing Observability tool or our SDK will be sent to Kubit's OTel endpoint.
The traces will be ingested in near real-time and available in Agent Activity, or you can check using skills

❯ /kubit-inspect traces in last 30 minutes

Enrichment batch job will run periodically to extract intent, sentiment and resolution from the traces
You can check the data pipeline status in Integration:

Data Masking

By default, Kubit automatically scrubs some sensitive data (e.g. credit card number, US social security number ) on our server side so they will never appear in our system.

Kubit OTel SDK provides helpers to mask any sensitive data you choose (set/replace, delete or mask). You can find the cookbook here.

Please note that Enrichment process (extracting intent, sentiment, resolution and persona) won't function if the LLM input and output are removed from the spans.

Manual Integration without SDK

If you already have OTel pipeline in your infrastructure, or just want to do things manually, you can simply setup an OTel Exporter to send your traces to Kubit. Please generate an API Key in Settings > Workspace > API Keys, then follow this cookbook.

Direct OTel integration may require some filtering, please contact us to set up and test properly.

Next steps

Find the traces from your application in Agent Activity, or analyze them in Dashboards or Reports.
Use Kubit Skills to get insights
Checkout Troubleshoot tips or Get Support if you run into troubles
Learn about advanced configuration like Field Mapping.