OpenLIT

What is OpenLIT?

OpenLIT is an open-source platform for observing and improving AI agents and LLM applications. Its SDKs produce OpenTelemetry traces for more than 90 LLMs and agent frameworks, so each LLM call, tool call, MCP request and retrieval can be inspected with token, cost and latency data. On top of tracing it adds evaluations using LLM-as-a-judge, programmatic and human feedback, runtime guardrails, a Prompt Hub for versioned prompts, a Vault for API keys, a Rule Engine, OpenGround for comparing models, and an AI copilot called Otter that turns questions into SQL. A CLI also tracks spend and usage for Claude Code, Cursor, Codex and OpenCode across a team. You can self-host OpenLIT for free under the Apache 2.0 license with Docker Compose or Helm, with ClickHouse for storage. An Enterprise build adds custom RBAC, audit logs and premium connectors; its pricing is available by contacting the OpenLIT team.

Use cases

• Trace LLM and agent calls with OpenTelemetry
• Track token spend by model and provider
• Score live traces with LLM-as-a-judge evals
• Monitor Claude Code and Cursor usage by team
• Version and deploy prompts without code changes
• Guard agents against prompt injection
• Monitor GPU utilization alongside LLM latency

Who it's for

• AI engineers
• Platform teams
• Engineering managers
• Self-hosters

Why use OpenLIT

• Built on OpenTelemetry, so traces are not locked into one vendor's format.
• Free to self-host under Apache 2.0, with the community edition covering observability, evals, guardrails and dashboards.
• Traces, evaluations, prompts and costs are linked, so a quality drop can be tied to a specific prompt version or model.
• Team-wide visibility into AI coding spend by developer, project, model and vendor.
• Prompts and keys can be changed centrally without redeploying the application.

OpenLIT features

• OpenTelemetry-native tracing for 90+ LLMs and agent frameworks
• Exceptions monitoring with stack traces and span context
• Token, cost and latency tracking with built-in model pricing
• Evaluations: LLM-as-a-judge, programmatic API and manual feedback
• Guardrails for prompt injection and topic restriction
• Prompt Hub with versioning, rollback and deployment to environments
• OpenGround for side-by-side model and prompt testing
• Vault for storing and rotating LLM API keys
• Coding-agent observability across Claude Code, Cursor, Codex and OpenCode
• Otter copilot for natural-language queries and dashboard generation
• GPU collector for GPU performance monitoring

OpenLIT integrations

• Claude Code
• Cursor
• Codex
• OpenCode
• OpenAI Agents SDK
• LangGraph
• CrewAI
• OpenTelemetry
• ClickHouse
• Datadog
• New Relic
• Slack
• PagerDuty
• Opsgenie
• Discord

How to get started with OpenLIT

1. Self-host OpenLIT and ClickHouse with Docker Compose or Helm.
2. Instrument your application with an OpenTelemetry-native SDK, or install the openlit CLI for coding agents.
3. Send traces to the OpenLIT OTLP receiver and view calls, costs and latency in dashboards.
4. Turn on evaluations and guardrails to score live traces and block risky inputs.
5. Use Prompt Hub, Vault, OpenGround and Otter to improve prompts and compare models.

OpenLIT limitations

• Self-hosting requires running the dashboard and a ClickHouse database yourself
• Custom RBAC, audit logs and premium connectors need the Enterprise build and a license
• Enterprise pricing is not published and is quoted on request

OpenLIT pricing

OpenLIT is open source, so you can use and self-host it for free.

OpenLIT plans

• Community

  Free

  self-hosted, Apache 2.0

  • AI observability and evaluations
  • Guardrails and dashboards
  • Prompt Hub, Vault and Rule Engine
  • Threshold alerts

• Enterprise

  Custom

  licensed per organisation

  • Custom role-based access control
  • Audit logs
  • Premium notification connectors
  • Datadog and New Relic data sources

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

OpenLIT alternatives

The main OpenLIT alternatives listed on Debuly are Langfuse, Arize Phoenix, Helicone, Arize AX, Opik and PromptLayer. The table compares pricing and platforms; order is A–Z.

1. 

   Arize AX

   Observability and evaluation platform for AI agents and LLM apps, with tracing, issue detection and experiments on a free plan.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Arize AX (official site)
2. 

   Arize Phoenix

   Open-source tool for tracing, evaluating and experimenting on LLM apps and agents, free to self-host or run locally.

   Developer platformsAutomation and agents
   Open sourceVisit Arize Phoenix (official site)
3. 

   Helicone

   Open-source LLM observability and AI gateway that logs requests, tracks cost and latency, and routes to 100+ models through one API.

   Developer platformsCoding
   FreemiumVisit Helicone (official site)
4. 

   Langfuse

   Open-source LLM observability and evaluation platform for tracing agents, managing prompts and scoring outputs.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Langfuse (official site)
5. 

   Opik

   Comet's open-source platform for tracing, evaluating and monitoring LLM applications and agents, self-hosted or in the cloud.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Opik (official site)
6. 

   PromptLayer

   Prompt management, evals and LLM observability so engineers and domain experts can ship prompt changes together.

   Developer platformsAutomation and agentsProductivity
   FreemiumVisit PromptLayer (official site)

OpenLIT FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from docs.openlit.io · fetched 24 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.