---
id: opentelemetry-instrumentation-replicate
version: "0.62.3"
license: Apache-2.0
license_treatment: permissive
maintenance: active
---

# opentelemetry-instrumentation-replicate — OpenTelemetry Replicate instrumentation

License: permissive · Maintenance: active · Downloads: 6.8M/mo

## What it is and what it does

This package hooks into the official Replicate Python library to automatically capture traces of your API calls. When you use Replicate to run models or generate images, the instrumentation records details like prompts, completions, and embeddings as OpenTelemetry span attributes, giving you visibility into what your application is sending and receiving.

By default, it logs all content to spans for debugging and quality evaluation. If that's a privacy concern—because prompts may contain sensitive user data—you can disable content logging by setting the environment variable TRACELOOP_TRACE_CONTENT to false, which reduces both privacy exposure and trace size. The package is designed to integrate seamlessly with existing OpenTelemetry observability stacks.

Use it for:

- Debug Replicate model outputs and prompts by inspecting span attributes in your tracing backend.
- Monitor LLM application behavior in production by exporting Replicate traces to Jaeger, Datadog, or similar.
- Evaluate prompt quality and model performance by analyzing trace data over time.
- Reduce trace verbosity and protect sensitive user data by disabling content logging in production.
- Correlate Replicate API calls with other application spans for end-to-end request tracing.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Adds OpenTelemetry tracing to Replicate API calls, capturing prompts, completions, and image generation requests as span attributes for observability and debugging.

Yes, if you use Replicate and already have OpenTelemetry infrastructure in place. The package is actively maintained, has no known vulnerabilities, and adds observability with minimal friction. Install it only if you need tracing; it requires Python 3.10+ and an OpenTelemetry SDK configured elsewhere in your application.

## Install

pip install opentelemetry-instrumentation-replicate
uv add opentelemetry-instrumentation-replicate
poetry add opentelemetry-instrumentation-replicate

## Installing opentelemetry-instrumentation-replicate

Before you install: Low install friction with a pure-Python wheel. Actively maintained as of 2026-08-10 with recent releases. Depends on four OpenTelemetry packages (api, instrumentation, semantic-conventions-ai, semantic-conventions), all standard observability infrastructure.

License in practice: Apache-2.0 permissive license allows use in commercial and open-source projects with minimal restrictions; attribution required.

Quickstart:

pip install opentelemetry-instrumentation-replicate

from opentelemetry.instrumentation.replicate import ReplicateInstrumentor

ReplicateInstrumentor().instrument()

Requires Python 3.10 or later. Assumes an OpenTelemetry SDK and exporter are already configured in your application.

Verify before relying:

- Whether this instrumentation works with all Replicate API endpoints or only specific model types.
- Performance overhead of tracing on Replicate API call latency.
- Compatibility with different OpenTelemetry exporter backends (Jaeger, Datadog, etc.).

## Package facts

- License: Apache-2.0 (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 6.8M/month (top 5,000 on PyPI)
- Known vulnerabilities: none known

## Tags

opentelemetry replicate tracing, replicate instrumentation, llm tracing replicate, opentelemetry ai instrumentation, replicate api observability, trace replicate prompts, replicate span attributes, observability, replicate-api, llm-tracing

[View on SkillFed](https://skillfed.io/packages/opentelemetry-instrumentation-replicate) · [View on PyPI](https://pypi.org/project/opentelemetry-instrumentation-replicate/)