LangChain Agent Monitoring That Scales With Your Fleet

Fluq provides a drop-in callback handler for LangChain and LangGraph. It automatically captures chains, tool calls, token usage, and errors without modifying your agent code. Add one callback and every LangChain run is traced and streamed to your dashboard.

Works alongside LangSmith or replaces it — your call. Fluq's event ingestion API is OpenTelemetry-compatible, so traces from Fluq integrate with your existing observability stack. No proprietary formats, no lock-in.

LangSmith shows you LangChain. Fluq shows you LangChain + CrewAI + AutoGen + custom agents, all in one view. Compare cost, latency, and error rates across frameworks. Understand how your LangChain agents perform relative to the rest of your fleet.

Unified alerting means you get notified about problems regardless of which framework built the agent. Fleet-level metrics like total throughput, aggregate cost, and cross-agent conflicts give you the full picture.

Install the SDK, import the callback handler, and pass it to your agent. Your dashboard populates immediately — no configuration, no schema definitions, no waiting. Free tier supports 50 agents with no credit card required.

pip install fluq-sdk
from fluq.integrations.langchain import FluqCallbackHandler
agent.run("Summarize the Q4 report", callbacks=[FluqCallbackHandler()])