September 2026 release is live Read More

Agent Cost & UsageIntegrations

Integrations

Choose the integration that matches how the agent application or platform sends telemetry to Mavvrik.

Mavvrik SDK

Use the SDK when the application can load a Mavvrik library at startup.

  1. SDK Integration — understand SDK behavior and instrumentation boundaries.

  2. Register an Agent — create the SDK agent identity and obtain credentials.

  3. Supported Agentic Stacks — confirm provider, framework, database, runtime, and language support.

  4. Continue with the Python SDK Quickstart or JavaScript / TypeScript SDK Quickstart.

eve

Use Connect eve for agents built with the eve framework.

eve exports its native OpenTelemetry telemetry directly to Mavvrik through agent/instrumentation.ts. The Mavvrik SDK is not required.

Langfuse

Use Connect Langfuse when agent telemetry is already collected in Langfuse.

No Mavvrik SDK registration is required before following the Langfuse connection flow.

LiteLLM Proxy

Use Connect LiteLLM when applications route model calls through a LiteLLM proxy.

This is a gateway integration. It is different from LiteLLM library instrumentation available in the Python SDK.

n8n

Use Connect n8n for self-hosted n8n workflows. Follow the n8n registration and connection flow on that page.

Select one primary telemetry path

Environment

Integration

Python agent application

Python SDK

JavaScript / TypeScript agent application

JavaScript / TypeScript SDK

eve agent application

eve

Existing agent telemetry in Langfuse

Langfuse

Model traffic routed through a LiteLLM proxy

LiteLLM Proxy

Self-hosted n8n workflow

n8n

Use one primary telemetry path for the same activity unless duplicate instrumentation is explicitly required. Instrumenting the same model calls through two paths can create duplicate records.