Open-CITE Bear Mascot

Open-CITE

Open Catalog of Intelligent Tools in the Enterprise

Discover and catalog every AI tool, model, and agent across your enterprise — from a single pane of glass.

Get Started View on GitHub
powered by

What is Open-CITE?

Open-CITE (pronounced "Open-Sight") is an open-source Python library, service, and application for discovering and cataloging AI/ML assets across your entire organization. It connects to your cloud platforms, collects traces from your AI agents, and builds a unified inventory of every tool, model, endpoint, and agent in your enterprise.

Run it as a GUI for interactive exploration, a headless API for automation, a Python library in your own code, or a containerized service in Docker and Kubernetes. Its plugin architecture means you only enable the discovery sources you need.

Key Features

Multi-Platform Discovery

Automatic discovery of AI/ML resources across Databricks, AWS, Google Cloud, Azure, and more — all from a single tool.

OpenTelemetry Native

Built-in OTLP receiver for real-time trace collection. Discover tools and models automatically from your AI agent telemetry.

Plugin Architecture

Extensible plugin system lets you add support for any platform. Enable only the discovery sources you need.

Lineage Graphs

Visualize relationships between tools, models, and agents with interactive network graphs showing how your AI assets connect.

Unified Schema Export

Export all discoveries in a standardized JSON format for downstream processing, governance tools, and compliance workflows.

Flexible Deployment

Run as a Python library, interactive GUI, headless REST API, or containerized service in Docker and Kubernetes.

See It in Action

Open-CITE Dashboard showing asset statistics, configured plugins, and lineage graph

The main dashboard shows discovery statistics, configured plugins with live status, and an interactive lineage graph.

Asset discovery grid showing discovered tools

The asset grid displays all discovered tools with type badges, discovery source, and mapping controls.

Lineage graph showing relationships between AI assets

Interactive lineage graphs reveal how tools, models, and agents connect across your organization.

Supported Platforms

Connect to the AI platforms you already use

Databricks
AWS Bedrock
AWS SageMaker
Google Cloud
Azure AI Foundry
OpenTelemetry
MCP
Splunk
Zscaler
Claude Code
Cursor
LangChain
MLflow
CrewAI
Cloudflare
OpenRouter

Getting Started

Up and running in under a minute

1

Install

# Create a virtual environment and install
python3 -m venv venv
source venv/bin/activate
pip install -e .
2

Launch the GUI

python -m open_cite.gui.app

# Open http://localhost:5000 in your browser
3

Add a Plugin

Click "+ Add Plugin" in the sidebar, select your platform (Databricks, AWS, Azure, Google Cloud, or OpenTelemetry), and enter your credentials.

4

Discover

Assets appear automatically as Open-CITE scans your environment. View tools, models, agents, and lineage in real time.

Or use Open-CITE as a Python library:

from open_cite import OpenCiteClient

client = OpenCiteClient(enable_otel=True)
tools = client.list_otel_tools()
client.export_to_json(filepath="discoveries.json")
Read the Full Documentation

Plugin Ecosystem

Enable only the discovery sources you need

  • Unity Catalog integration — catalogs, schemas, tables, and volumes
  • MLflow trace search and retrieval for observability
  • AI Gateway usage tracking and Genie conversation analysis
  • Data lineage and metadata management
View Documentation →
  • Discovers foundation models, custom models, and invocations
  • Model inference logging and usage tracking
  • Automatic credential discovery via AWS profiles
View Documentation →
  • Discovers endpoints, models, model packages, and training jobs
  • Model deployment tracking across regions
  • Integration with SageMaker Studio notebooks
View Documentation →
  • Vertex AI model and endpoint discovery
  • Generative AI model listing (Gemini, PaLM, and more)
  • MCP server discovery via labels and port scanning
View Documentation →
  • Discovers AI resources, model deployments, and projects
  • Agent and tool discovery via Assistants API
  • Trace discovery via Log Analytics integration
View Documentation →
  • Built-in OTLP/HTTP and gRPC receiver for trace collection
  • Automatic tool and model discovery from trace data
  • Works with any LLM provider (OpenAI, Anthropic, etc.)
  • OpenRouter broadcast support for zero-code integration
View Documentation →
  • Trace-based MCP server discovery
  • MCP tool and resource cataloging
  • Usage pattern analysis across MCP servers
View Documentation →

Open Source on GitHub

Browse the code, file issues, and contribute to the project.

View Repository

Enterprise AI Governance

Want to govern and monitor your AI assets? Try LangGuard's AI Control Plane.

Visit LangGuard.AI