Neptune MCP
Track ML experiments with AI via MCP. Query Neptune runs, compare models, manage metadata, and analyze experiment results.
What is Neptune MCP?
Neptune MCP integrates Neptune.ai's experiment tracking platform with AI through MCP. Neptune provides flexible metadata storage for ML experiments, enabling teams to log, display, organize, and compare any metadata.
Flexible Metadata Intelligence
Neptune's flexible metadata model means AI models can query any logged data — metrics, parameters, artifacts, tags, and custom metadata — providing comprehensive experiment insights.
Configuration
{
"mcpServers": {
"neptune": {
"command": "npx",
"args": ["neptune-mcp"],
"env": {
"NEPTUNE_API_TOKEN": "your_api_token"
}
}
}
}Use Cases
Neptune MCP serves ML teams needing flexible experiment tracking, researchers comparing across hundreds of runs, and organizations requiring comprehensive ML metadata management.
Key Features
- Query experiment runs and metadata
- Compare model performance metrics
- Track dataset versions and lineage
- Monitor training hardware utilization
- Manage project-level dashboards
- Analyze experiment parameter importance
Similar MCP Servers
View all →
Everything Claude Code
The agent harness performance optimization system.
Mcp For Beginners
This open-source curriculum introduces the fundamentals of MCP.
DesktopCommanderMCP
MCP server for Claude with terminal control and file search.
Docker Hub MCP
Official MCP server to interact with Docker Hub.