Alternatives to Langfuse

MLflow is the open source AI engineering platform for agent and LLM observability, evaluation, and prompt management.

Open-source tracing, evaluation, and prompt iteration for AI agents — self-host it on your own infrastructure with no per-span bill.

Open-source AI evaluation and observability for LLMs, RAG, agents, and predictive ML models.

Open-source LLM observability, prompt management, and agent evals in one MIT-licensed platform.

OpenLIT is an open-source, OpenTelemetry-native platform for LLM tracing, evaluation, and prompt management.

OpenJudge is an open-source AI evaluation framework with 50+ production-grade graders for agents, multimodal models, code, and math.

Arize Phoenix is open-source LLM observability and evals that trace every agent step so you can ship reliable AI agents.

Collaborative LLM development platform for prompt management, evaluation, and production monitoring

Open-source framework to replace vibe checks with reproducible, LLM-driven evaluation loops for RAG and agents.

LangSmith is an AI agent observability platform for tracing, monitoring, and evaluating LLM apps and long-running agents.

Opik, Comet's open-source LLM observability and eval platform, turns agent traces into root-cause groupings and git-committed code fixes

Open-source OpenTelemetry observability for Python LLM apps

Open-source, OpenTelemetry-native agent evaluation and tracing that finds where your agent fails.

Open-source workspace for building, evaluating, and deploying AI agents you talk to in Slack, Telegram, or a browser chat

Maxim AI simulates, evaluates, and observes AI agents in one workspace, with an open-source gateway (Bifrost) for routing and governance.

AI observability platform that turns offline evals into production guardrails for AI agents and RAG systems.

Open-source AI observability for agent tracing, LLM-as-a-judge evals, and coding agent cost tracking

TruEra brings ML monitoring, testing, and AI quality management into Snowflake for production model observability.

Raindrop is agent observability that finds silent failures in production AI agents, investigates the root cause, and simulates the fix before you merge.

Open-source AI agent observability that catches failures automatically and helps you fix them.

OpenTelemetry-native observability with an AI SRE that investigates incidents and opens fix PRs.

AI observability and eval-engineering platform that turns offline evals into live production guardrails.

Enterprise LLM evaluation, observability, and AI red teaming that standardizes quality across every team.

Discontinued: WhyLabs ceased operations in 2025 and open-sourced its platform; whylogs and langkit live on as open source.

Agent observability that traces every AI run, scores quality with evals, and surfaces production patterns you didn't know to look for.

Open-source prompt management and observability for LLMs

OpenCompass (司南) benchmarks LLMs, VLMs, and AI4S models with 100+ open evaluation datasets

Metoro is an AI SRE agent for Kubernetes observability, using eBPF telemetry to detect, root-cause, and fix production issues.

Free open-source AI usage tracking & analytics for developers

Open-source observability and self-repair for AI agents, turning production failures into validated, reusable rules.