AI Agent Hub
Back to plugins
🧰

dsh-plugin-langfuse

Web Tools Updated 2026.08.21

Run the following command in DeepSeek Harness:

dsh plugin install linyp/dsh-plugin-langfuse

Paste the following prompt into your AI chat to install this plugin:

You can install this plugin by running `dsh plugin install linyp/dsh-plugin-langfuse`, and access the full source code at https://github.com/linyp/dsh-plugin-langfuse .

About this plugin

Building complex LLM agents often involves navigating the black box of multi-turn conversations, tool executions, and context compactions. dsh-plugin-langfuse brings native Langfuse observability to DeepSeek Harness, transforming every runtime interaction into standardized OpenTelemetry traces and making your agent's reasoning and execution fully transparent.

Deeply integrated with the Harness telemetry seam, this plugin accurately captures model generations and tool spans while gracefully handling complex scenarios like retries, interruptions, context compactions, and subagent lineages. It also maps canonical user feedback to Langfuse Scores and offers flexible content redaction policies to ensure sensitive data is sanitized before export.

Whether you are an engineer building AI applications with DeepSeek Harness or a team needing to monitor token costs, debug agent toolchains, and evaluate model performance, this plugin serves as the perfect bridge between your local runtime and cloud monitoring, helping you effortlessly establish a production-grade LLM observability stack.

Use Cases

  • Trace multi-turn conversations and tool call chains to debug agent errors.
  • Monitor token usage and model latency to optimize LLM application costs.
  • Collect user feedback scores to evaluate and iterate on model generation quality.

Best For

  • AI engineers developing complex LLM agents.
  • Operations teams monitoring production model performance and costs.
  • Full-stack developers focused on improving LLM application observability.