Analysis updated 2026-05-18
Send OpenClaw agent and LLM call traces to Langfuse for monitoring.
Track token usage and prompt and response content across agent runs.
Debug which tools and agents ran during a specific AI request.
Route infrastructure metrics like gateway error rates to a separate monitoring pipeline.
| aliyun/openclaw-exporter-to-langfuse | anousss007/ng-blatui | blockedpath/pi-xai-oauth | |
|---|---|---|---|
| Stars | 11 | 11 | 11 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | moderate | moderate | easy |
| Complexity | 4/5 | 3/5 | 2/5 |
| Audience | ops devops | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Installation is version-matched to your OpenClaw release via a compatibility matrix, typically run by an AI agent.
openclaw-exporter-to-langfuse is a plugin that collects observability data from OpenClaw, an AI gateway used for routing and managing LLM requests, and sends it to Langfuse, a platform for monitoring and evaluating AI applications. When you run AI agents that call language models and tools, it can be hard to understand what actually happened inside a given request: which agents ran, which tools were called, how many tokens were used, and what the prompts and responses looked like. This plugin hooks into OpenClaw and captures that information as structured spans, labeled records for each step such as agent invocations, LLM calls, tool executions, and session start and end events. Those spans follow the OpenTelemetry GenAI semantic conventions, a standard format for AI observability data, and are forwarded to Langfuse where you can view call chains, token usage, and prompt and response content. Separately and optionally, the plugin also supports routing infrastructure level metrics, gateway request rates, error rates, system logs, through a different pipeline to a data store and visualization tool of your choice. The README notes that on Alibaba Cloud, a managed offering called Agent-lens builds on top of this same infrastructure. Installation is handled by a shell script and managed by an AI agent that checks your OpenClaw version, looks up the compatible exporter version from a JSON compatibility matrix, and runs the appropriate installer. The plugin is written in TypeScript and licensed under MIT.
A plugin that captures AI agent, LLM, and tool call tracing data from OpenClaw and sends it to Langfuse for monitoring.
Mainly TypeScript. The stack also includes TypeScript, OpenTelemetry, Langfuse.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly ops devops.
This repo across BitVibe Labs
Verify against the repo before relying on details.