OpenLIT
Coding assistants and agents · OpenLIT · Open source project
Open-source observability and evaluation platform for AI agents and coding agents using OpenTelemetry. docs.openlit.io ↗
Facts
DescriptionOpenLIT is an open-source platform that traces LLM and agent calls via OpenTelemetry, tracks cost and token usage, runs automated LLM-as-a-judge evaluations, and manages versioned prompts. It also offers observability for coding agents like Claude Code, Cursor, and Codex, tracking spend and productivity per developer. It ships with Otter, a built-in AI copilot for querying telemetry in natural language, and can be self-hosted via Docker Compose or Helm.
DeploymentOpen sourceOn-premSaaS
PricingFree
Pricing detailOpen source and self-hostable; no commercial pricing stated.
Target buyersengineeringdata
IntegrationsOpenTelemetryClaude CodeCursorCodexCrewAILangGraphDocker ComposeHelmMintlify
Models usedClaude CodeCursorCodex
Features · 15
llm tracingagent observabilitycoding agent spend trackingprompt versioningprompt huba/b model testingllm-as-a-judge evalsci/cd regression testingcost trackingtoken analyticssecret redactionapi key vaultnatural language telemetry queriesai dashboard generationgpu utilization monitoring
Dossier · model-written summary
Marketing-heavy docs page for an open-source LLM observability platform; it does much more than the believed category, spanning tracing, evals, prompt management, and coding agent spend tracking. No named customers, pricing, or compliance certifications are stated. The OpenTelemetry-native, self-hostable design is credible for regulated enterprises, but maturity and vendor details are unclear. Cod
Timeline · 1 event
Same category · Coding assistants and agents