OpenInference
OpenTelemetry Instrumentation for AI Observability
What is OpenInference?
OpenInference is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to opentelemetry instrumentation for ai observability
OpenTelemetry Instrumentation for AI Observability
This server falls under the Monitoring & Observability category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- OpenTelemetry Instrumentation for AI Observability
Use Cases
Maintainer
Works with
Installation
NPM
npx -y @arizeai/openinference-instrumentation-mcpManual Installation
npx -y @arizeai/openinference-instrumentation-mcpConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use OpenInference
OpenInference is an open-source observability framework built on top of OpenTelemetry that adds AI-specific tracing conventions for LLM applications, agents, and MCP interactions. The `openinference-instrumentation-mcp` package instruments Model Context Protocol client-server communications, capturing span-level traces of every tool call, protocol message, and response so teams can debug, monitor, and optimize their AI pipelines. It integrates natively with Arize Phoenix, Arize AX, and any OpenTelemetry-compatible observability backend, making it the standard instrumentation layer for production LLM applications built on LangChain, LlamaIndex, LangGraph, and Haystack.
Prerequisites
- Python 3.8+ with pip
- An OpenTelemetry-compatible observability backend: Arize Phoenix (open source), Arize AX, Jaeger, or any OTLP-compatible collector
- An existing Python application using MCP client libraries
- Optional: Node.js 18+ for the npm package (@arizeai/openinference-instrumentation-mcp)
Install the Python MCP instrumentation package
Install openinference-instrumentation-mcp via pip. This adds OpenTelemetry spans for all MCP protocol interactions in your Python application.
pip install openinference-instrumentation-mcpOr install the npm package for Node.js projects
For TypeScript or JavaScript MCP applications, install the npm package instead.
npm install @arizeai/openinference-instrumentation-mcpSet up an OpenTelemetry tracer provider
Configure your application with an OTLP exporter pointing to your observability backend. Arize Phoenix can be run locally for development.
from opentelemetry import trace
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
from opentelemetry.sdk.trace.export import SimpleSpanProcessor
provider = TracerProvider()
exporter = OTLPSpanExporter(endpoint="http://localhost:6006/v1/traces")
provider.add_span_processor(SimpleSpanProcessor(exporter))
trace.set_tracer_provider(provider)Instrument your MCP client
Apply the MCP instrumentor to automatically capture traces for all MCP tool calls and protocol messages without modifying individual call sites.
from openinference.instrumentation.mcp import MCPInstrumentor
MCPInstrumentor().instrument()Run your application and view traces
Start your MCP-powered application. All client-server interactions will be traced and visible in your configured observability backend (e.g. Arize Phoenix at http://localhost:6006).
# Start Phoenix locally
pip install arize-phoenix
python -m phoenix.server.main serve
# Then run your instrumented applicationOpenInference Examples
Client configuration
MCP server configuration for the OpenInference instrumentation server when used as an MCP server component.
{
"mcpServers": {
"openinference": {
"command": "npx",
"args": ["-y", "@arizeai/openinference-instrumentation-mcp"]
}
}
}Prompts to try
Once OpenInference instrumentation is active in your AI application, you can query trace data and analyze behavior.
- "Show me all MCP tool calls made in the last hour that took more than 2 seconds"
- "Which MCP tools are being called most frequently in my agent pipeline?"
- "Find any MCP interactions that returned errors in today's traces"
- "Compare the latency of tool calls between my local and remote MCP servers"Troubleshooting OpenInference
No traces appearing in the observability backend
Ensure MCPInstrumentor().instrument() is called before any MCP client connections are established. Verify the OTLP endpoint URL is correct and the observability backend is running and reachable.
Sensitive data appearing in traces
OpenInference supports OPENINFERENCE_HIDE_* environment variables to suppress or mask sensitive fields. Set OPENINFERENCE_HIDE_INPUT_TEXT=true or OPENINFERENCE_HIDE_OUTPUT_TEXT=true to redact content from spans.
Import error: 'openinference.instrumentation.mcp' not found
Ensure you installed the correct package: 'pip install openinference-instrumentation-mcp'. The package name uses hyphens but the Python import uses underscores.
Frequently Asked Questions about OpenInference
What is OpenInference?
OpenInference is a Model Context Protocol (MCP) server that opentelemetry instrumentation for ai observability It connects AI assistants to external tools and data sources through a standardized interface.
How do I install OpenInference?
Install via npm with the command: npx -y @arizeai/openinference-instrumentation-mcp. Then add the server configuration to your AI client's JSON config file (e.g., claude_desktop_config.json or .cursor/mcp.json).
Which AI clients work with OpenInference?
OpenInference works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is OpenInference free to use?
Yes, OpenInference is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.
OpenInference Alternatives — Similar Monitoring & Observability Servers
Looking for alternatives to OpenInference? Here are other popular monitoring & observability servers you can use with Claude, Cursor, and VS Code.
Netdata
★ 78.9kReal-time infrastructure monitoring with metrics, logs, alerts, and ML-based anomaly detection.
Kubeshark
★ 11.9keBPF-powered network observability for Kubernetes. Indexes L4/L7 traffic with full K8s context, decrypts TLS without keys. Queryable by AI agents via MCP and humans via dashboard.
Mission Control
★ 4.9kSelf-hosted AI agent orchestration platform: dispatch tasks, run multi-agent workflows, monitor spend, and govern operations from one mission control dashboard.
Grafana
★ 3.0kThis MCP server enables natural-language querying of Grafana logs by automatically detecting log sources and service labels. It provides read-only access to log data with intelligent caching for efficient repeat queries.
Sentrux
★ 2.4kReal-time architectural sensor that helps AI agents close the feedback loop, enabling recursive self-improvement of code quality. Pure Rust.
Thinkwatch
★ 967Enterprise AI bastion host for secure AI API and MCP access, with unified proxying, RBAC, audit logs, rate limiting, and cost tracking across OpenAI, Anthropic, Gemini, and self-hosted LLMs.
Browse More Monitoring & Observability MCP Servers
Explore all monitoring & observability servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up OpenInference in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use OpenInference?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.