# OpenLIT vs OpenLLMetry

> OpenLLMetry (by Traceloop) is an instrumentation library that emits OpenTelemetry spans for LLM apps. OpenLIT is a full self-hosted platform (UI, evaluations, Prompt Hub, agents, Vault, GPU) with its own SDKs that also ingest and export OTLP.

- HTML: https://openlit.io/compare/openlit-vs-openllmetry
- Markdown: https://openlit.io/compare/openlit-vs-openllmetry.md

Compare OpenLIT and OpenLLMetry. OpenLLMetry is an OpenTelemetry instrumentation library. OpenLIT is a full self-hosted AI engineering platform plus SDKs that also speak OTLP.

## Feature comparison

### What it is

| Feature | OpenLIT | OpenLLMetry |
| --- | --- | --- |
| Category | Full AI engineering platform + SDK | Instrumentation library (OTLP exporter) |
| OpenTelemetry-native | Yes | Yes |
| Open Source | Apache 2.0 | Apache 2.0 (SDK) |
| Self-hosted product UI | Yes | No |
| Requires a separate backend | Optional (ships with platform stack) | Yes |

### LLM Monitoring

| Feature | OpenLIT | OpenLLMetry |
| --- | --- | --- |
| Auto-instrument LLM calls | Yes | Yes |
| Token / cost attributes on spans | Yes | Yes |
| Built-in traces UI | Yes | No |
| Built-in dashboards | Yes | No |

### Platform features

| Feature | OpenLIT | OpenLLMetry |
| --- | --- | --- |
| Evaluations (LLM-as-a-judge) | Yes | No |
| Prompt Hub (versioning) | Yes | No |
| Secrets Vault | Yes | No |
| AI agent monitoring UI | Yes | No |
| GPU collector | Yes | No |

### Interoperability

| Feature | OpenLIT | OpenLLMetry |
| --- | --- | --- |
| Can send OTLP to OpenLIT | N/A (is the platform) | Yes |
| Works with Grafana / any OTLP backend | Yes | Yes |

## When to choose OpenLIT

- You want a self-hosted UI for traces, evals, prompts, agents, and costs, not only an SDK
- You need Prompt Hub, LLM-as-a-judge, Vault, or GPU monitoring out of the box
- You prefer one Apache-2.0 platform that includes instrumentation and the backend
- You want to accept OTLP from OpenLLMetry or other OTel sources into OpenLIT

## When to choose OpenLLMetry

- You only need an instrumentation library and already have an OTLP backend
- You are standardising on Traceloop/OpenLLMetry instrumentation across services
- You do not need a prompt, eval, or agent product UI

## Related

- All comparisons: https://openlit.io/compare.md
- Pricing: https://openlit.io/pricing.md
- Docs: https://docs.openlit.io/latest/openlit/quickstart-ai-observability