# OpenLIT review

> OpenLIT is an open-source, self-hosted platform for tracing, evaluating, guarding, and improving AI agents and coding agents on OpenTelemetry.

Devtoolwatch Board, the week of Oct 25–31, 2026. Canonical page: https://index.devtoolwatch.com/tools/openlit/

## TL;DR

OpenLIT is an open-source agent harness engineering platform with OpenTelemetry-native tracing, evals, guardrails, prompt management, and cost and GPU monitoring. It is for teams running AI agents or coding agents in production. OpenLIT is free to self-host under Apache 2.0, follows OpenTelemetry GenAI semantic conventions, and can export to other OTLP backends.

## OpenLIT at a glance

- **Category:** [Evals and observability](https://index.devtoolwatch.com/categories/evals-and-observability/)
- **Maintained by:** OpenLIT (organization, [@openlit](https://github.com/openlit) on GitHub)
- **License:** Apache 2.0
- **Language:** TypeScript, Python, Go
- **Repository:** [github.com/openlit/openlit](https://github.com/openlit/openlit)
- **Website:** [docs.openlit.io](https://docs.openlit.io)
- **npm package:** [openlit](https://www.npmjs.com/package/openlit) 1.15.0
- **PyPI package:** [openlit](https://pypi.org/project/openlit/) 1.45.0
- **Latest release:** [otel-gpu-collector-1.0.0](https://github.com/openlit/openlit/releases/tag/otel-gpu-collector-1.0.0) on GitHub, Oct 13, 2026
- **First published:** Jan 23, 2024

## What OpenLIT does

- OpenLIT traces LLM calls, tool calls, MCP requests, retrieval, and agent steps as OpenTelemetry traces and metrics.
- OpenLIT runs evaluations on real traces using LLM-as-a-judge, programmatic checks, human feedback, and CI gates.
- OpenLIT SDK guardrails detect and block prompt injection, jailbreak attempts, sensitive topics, and off-topic requests at runtime.
- OpenLIT captures coding-agent sessions from Claude Code, Cursor, Codex, and OpenCode, including tool calls, token usage, cost, and code impact.
- OpenLIT tracks LLM cost with custom pricing and monitors NVIDIA, AMD, and Intel GPUs used for inference.
- OpenLIT Prompt Hub stores versioned prompts, and Vault stores API keys retrieved at runtime via SDK or API.
- OpenLIT OpenGround runs the same prompt across multiple providers to compare cost, latency, and output quality.

### How it works

OpenLIT runs in your infrastructure with Docker or Kubernetes. Your app sends telemetry over OTLP to the OpenLIT receiver on ports 4317 (gRPC) or 4318 (HTTP). Data is stored in ClickHouse and shown in the OpenLIT dashboard. Python, TypeScript, and Go SDKs auto-instrument supported providers and frameworks. You can place your own OpenTelemetry Collector in front to fan out to other backends.

## How to install OpenLIT

**Python SDK with pip**

```
pip install openlit
```

**TypeScript SDK with npm**

```
npm install openlit
```

### Requirements

- Python 3.9 or later (below 4.0) for the Python SDK.
- Node.js 18.0.0 or later for the TypeScript SDK.
- Docker or Kubernetes to self-host the OpenLIT platform.

### First steps

1. Start OpenLIT with Docker Compose and open http://127.0.0.1:3000 in a browser.
2. Install the Python or TypeScript SDK and call openlit.init() in your application.
3. Point telemetry at the OTLP endpoint by setting OTEL_EXPORTER_OTLP_ENDPOINT to http://127.0.0.1:4318.
4. For coding agents, install the CLI, then run openlit coding install --vendor=all and check with openlit doctor.

## Who OpenLIT is for

- Teams running AI agents in production who want traces, evals, and guardrails in one self-hosted tool.
- Developers who want to see what Claude Code, Cursor, Codex, or OpenCode did during a session.
- Teams with an existing OpenTelemetry setup who want GenAI telemetry without a proprietary format.
- Teams serving LLMs on their own NVIDIA, AMD, or Intel GPUs.

### Limitations

- OpenLIT is not an agent framework or harness runtime and does not run your agent loop.
- OpenLIT self-hosting means you operate the deployment, including its ClickHouse storage.
- OpenLIT's TypeScript SDK package on npm lists an ISC license, while the repository and Python package list Apache-2.0.

## Usage data (the week of Oct 25–31, 2026)

- Board rank: #249 overall, #12 in Evals and observability, momentum score 38
- Downloads that week: 88.1k (+11% vs 4-week average)
- New GitHub stars per week (last 12 weeks): Aug 9: 16; Aug 16: 22; Aug 23: 15; Aug 30: 16; Sep 6: 18; Sep 13: 13; Sep 20: 26; Sep 27: 19; Oct 4: 26; Oct 11: 20; Oct 18: 28; Oct 25: 16

## OpenLIT alternatives

| Rank | Tool | Stars | Downloads/wk |
|---|---|---|---|
| #1 | [SkillSpector](https://index.devtoolwatch.com/tools/nvidia-skillspector/) | 23.7k | – |
| #2 | [Judgeval](https://index.devtoolwatch.com/tools/judgmentlabs-judgeval/) | 1.1k | 80.2k |
| #3 | [A.I.G (AI-Infra-Guard)](https://index.devtoolwatch.com/tools/tencent-ai-infra-guard/) | 7.2k | – |
| #4 | [Langfuse](https://index.devtoolwatch.com/tools/langfuse/) | 36.8k | 9.8M |
| #5 | [Opik](https://index.devtoolwatch.com/tools/opik/) | 23.1k | 709.1k |
| #6 | [promptfoo](https://index.devtoolwatch.com/tools/promptfoo/) | 26.7k | 1.2M |

## OpenLIT FAQ

### Is OpenLIT free and open source?

Yes. OpenLIT's source code is on GitHub (openlit/openlit) under the Apache 2.0 license, so you can use it for free.

### How do I install OpenLIT?

Run `pip install openlit`. It needs Node.js >=18.0.0. The install section above has the other options from the project's README.

### Is OpenLIT still maintained?

Yes. The openlit/openlit repository was last updated on Oct 21, 2026, and the latest release, otel-gpu-collector-1.0.0 on GitHub, came out on Oct 13, 2026.

### How popular is OpenLIT?

OpenLIT has 2.9k GitHub stars as of Nov 1, 2026. It gained 16 new stars in the week of Oct 25–31, 2026, against a 12-week average of 20 a week. Its main package had about 88.1k downloads that week, +11% against its own 4-week average. That made it #12 among LLM evals and observability tools on the Devtoolwatch Board.

### OpenLIT vs Langfuse: how do they compare?

OpenLIT and Langfuse are both tracked by Devtoolwatch as LLM evals and observability tools. OpenLIT has 2.9k GitHub stars and Langfuse has 36.8k. In the week of Oct 25–31, 2026, OpenLIT's main package had 88.1k downloads and Langfuse's had 9.8M. Langfuse is an open source LLM engineering platform for tracing, evaluating, prompt management, and debugging LLM applications and agents. [Langfuse review](https://index.devtoolwatch.com/tools/langfuse/)

### OpenLIT vs OpenLLMetry: how do they compare?

OpenLIT and OpenLLMetry are both tracked by Devtoolwatch as LLM evals and observability tools. OpenLIT has 2.9k GitHub stars and OpenLLMetry has 7.5k. OpenLLMetry is an open-source set of OpenTelemetry extensions that adds observability to LLM applications, with instrumentations for LLM providers, vector DBs and frameworks. [OpenLLMetry review](https://index.devtoolwatch.com/tools/openllmetry/)

### What are the best alternatives to OpenLIT?

Other LLM evals and observability tools on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: SkillSpector (#1), Judgeval (#2), A.I.G (AI-Infra-Guard) (#3), Langfuse (#4), Opik (#5). The alternatives table above has their numbers.

### Is OpenLIT built on OpenTelemetry?

Yes. OpenLIT is OpenTelemetry-native. Traces and metrics follow the OpenTelemetry GenAI semantic conventions (gen_ai.*), and telemetry can flow through your own Collector to other OTel backends such as Datadog, Grafana, or Honeycomb.

### Does OpenLIT replace my agent framework?

No. OpenLIT does not run your agent loop. OpenLIT instruments the harness you already use, such as LangGraph, CrewAI, or the OpenAI Agents SDK, through OpenTelemetry.

### Does OpenLIT work with Claude Code, Codex, and Cursor?

Yes. The openlit CLI ingests each coding agent's hook events and maps them to OpenTelemetry gen_ai.* conventions, with no SDK and no code changes.

### Do I need an OpenTelemetry Collector to use OpenLIT?

No. OpenLIT listens for OTLP directly on port 4317 (gRPC) and 4318 (HTTP). You can still put your own Collector in front for fan-out or processing.

### What does an OpenLIT guardrail detect?

OpenLIT SDK guardrails detect prompt injection and jailbreak attempts, sensitive topics, and topic restriction violations. Guardrail checks are traced with OpenTelemetry, so you can see when and why a request was blocked.

## Sources

1. [openlit/openlit on GitHub](https://github.com/openlit/openlit)
2. [OpenLIT README](https://github.com/openlit/openlit/blob/main/README.md)
3. [openlit on npm](https://www.npmjs.com/package/openlit)
4. [openlit on PyPI](https://pypi.org/project/openlit/)
5. [OpenLIT website](https://docs.openlit.io)
6. [Release otel-gpu-collector-1.0.0 (Oct 13, 2026)](https://github.com/openlit/openlit/releases/tag/otel-gpu-collector-1.0.0)
7. [Release openlit-2.1.0 (Sep 10, 2026)](https://github.com/openlit/openlit/releases/tag/openlit-2.1.0)
8. [Release openlit-2.0.0 (Aug 28, 2026)](https://github.com/openlit/openlit/releases/tag/openlit-2.0.0)
9. [Hacker News: Show HN: OpenLIT – Open-Source LLM Observability with OpenTelemetry](https://news.ycombinator.com/item?id=40167461)
10. [Devtoolwatch Board, the week of Oct 25–31, 2026](https://index.devtoolwatch.com/)