Build LazyAnalysis applications with your AI.
Connect Claude, ChatGPT or another MCP-compatible assistant and generate downloadable .lazy graphs and dashboards.
What this connector does
LazyAnalysis is a computational application composer, not an agent-to-agent workflow framework. AI assistants discover the actual native node catalog, assemble optimized computation and controls into typed graphs, and produce validated .lazy files and accompanying dashboards. Use native nodes for performance, Formula nodes for expressions, Python nodes for supported custom logic, and nested parameterized Super Nodes to organize reusable subsystems.
Connect Claude
In Claude, open Customize or Settings → Connectors → Add custom connector. Set the name to LazyAnalysis and the remote MCP URL to https://mcp.lazyanalysis.it/mcp. The public server currently requires no OAuth sign-in or custom request header. Enable its tools in a conversation. Availability of custom connectors depends on your Claude plan or organization settings.
Connect ChatGPT or another MCP client
If your ChatGPT plan or workspace permits custom MCP connections, add a remote Streamable HTTP MCP server with URL https://mcp.lazyanalysis.it/mcp, then authorize use of its graph-building tools. Alternatively, use any MCP-compatible client that supports remote HTTPS servers. MCP Registry identity: it.lazyanalysis/lazyanalysis. Tool access and directory availability depend on the AI provider.
Compose your first solution
Ask your AI: “Using LazyAnalysis MCP, create a complete vibration-monitoring .lazy application with a synthetic signal generator, FFT analysis, configurable thresholds, plots, status indicators, and a professional dashboard. Use named parameterized Super Nodes for reusable processing. Discover the real node catalog, validate the result, and give me its download link.” Adapt the request to your needs and verify the produced graph before using it operationally.
Download, open, and run
The assistant calls the MCP create_graph tool and receives a downloadable .lazy link. Save the file, open it inside LazyAnalysis using File > Open, inspect its graph, data sources and dashboard, then execute it locally or in your own authorized application session. The public MCP server does not execute graphs, access your hardware, or inspect runtime values.
What is available and what is not
The public server provides 13 MCP tools for capability discovery, node types and settings, appearance and widget styles, solution patterns, graph generation, validation, and graph-to-spec conversion. Exact Formula/Python node settings, dependencies, available protocols, GPU features and hardware access must be checked against the target runtime and catalog. Graph validation is structural, not a guarantee of correct results or safe equipment operation.
Data and support
Graph files generated for download are stored on the MCP server for up to seven days (and may be removed earlier under storage limits). Anyone with a download URL may access that file while it remains available; do not submit confidential data without authorization. Read /en/privacy/ for data handling or email support@lazyanalysis.it for MCP assistance.