Skip to main content

​
Overview

This guide shows you how to integrate Latitude Telemetry into an application that uses Mastra. Mastra has built-in observability support via @mastra/observability and @mastra/otel-exporter. It emits standard OpenTelemetry gen_ai.* spans, so you can point Mastra’s OTel exporter directly at Latitude’s OTLP ingestion endpoint — no Latitude SDK required.
You’ll keep using Mastra exactly as you do today. The exporter simply sends your traces to Latitude alongside any other observability backend.
The Mastra integration is TypeScript only.

​
Requirements

  • A Latitude account and API key
  • A Latitude project slug
  • A Node.js project that uses Mastra (@mastra/core)

​
Steps

1

Install

npm install @mastra/observability @mastra/otel-exporter @opentelemetry/exporter-trace-otlp-proto
2

Configure observability

Create an OtelExporter with a custom provider pointed at Latitude’s OTLP endpoint, then pass it to Mastra’s Observability config:
import { Mastra } from "@mastra/core"
import { Agent } from "@mastra/core/agent"
import { Observability } from "@mastra/observability"
import { OtelExporter } from "@mastra/otel-exporter"

const otelExporter = new OtelExporter({
  provider: {
    custom: {
      endpoint: "https://ingest.latitude.so/v1/traces",
      protocol: "http/protobuf",
      headers: {
        Authorization: `Bearer ${process.env.LATITUDE_API_KEY!}`,
        "X-Latitude-Project": process.env.LATITUDE_PROJECT_SLUG!,
      },
    },
  },
})

const agent = new Agent({
  id: "my-agent",
  name: "My Agent",
  model: { provider: "OPEN_AI", name: "gpt-4o" },
  instructions: "You are a helpful assistant.",
})

const mastra = new Mastra({
  agents: { "my-agent": agent },
  observability: new Observability({
    configs: {
      otel: {
        serviceName: "my-mastra-app",
        exporters: [otelExporter],
      },
    },
  }),
})

​
Seeing Your Traces

Once connected, traces appear automatically in Latitude:
  1. Open your project in the Latitude dashboard
  2. Each execution shows input/output messages, model, token usage, latency, and errors
  3. Mastra agent runs, tool executions, and workflow steps appear as child spans