Cheburnexus analyzes your code entirely on your own machine. Your source code is never uploaded, transmitted, or sent to any server operated by us or by any third party. There is no telemetry.
The analysis engine and the MCP server are native executables that run as local processes on your computer. They read your project's source files from disk and write their analysis output (.cheburnexus/out/) back to disk, inside your project folder.
The MCP server listens only on 127.0.0.1 (your own machine's loopback interface) — it is never reachable from the network.
JetBrains AI Assistant calls the MCP server locally to read the analysis results. What AI Assistant does with the data it retrieves is governed by JetBrains' own privacy policy, not this one — we do not see or control that traffic.
Engine download (one-time). On first activation, the plugin downloads the analysis engine (~250 MB) from our public GitHub Releases (github.com/Oberonru/cheburnexus-releases). This is a plain file download over HTTPS, verified against a fixed SHA-256 checksum before use. No project data is sent as part of this request — it is a one-way download only.
License/subscription page. The "Get a Subscription" action opens your subscription page in your default browser. That page is operated separately and has its own privacy terms.
claims.json) on your machine, not validated against a remote server.claims.json (your license/entitlement passport, if you have a paid plan) and usage.json (a local counter of tokens saved, used only to enforce your plan's limits) — both stored in your OS's local application-data folder, never transmitted anywhere.
.cheburnexus/out/ inside your analyzed project — the analysis output (dependency graph, metrics, findings) AI Assistant reads. This is ordinary local project data; treat it like any other build artifact (you may .gitignore it).
A license key, if you enter one, is stored using the IDE's built-in PasswordSafe (the same secure-storage mechanism the IDE uses for other credentials), not in a plain config file.
The plugin also writes a Claude-Desktop-format MCP registration file (.ai/mcp/mcp.json in your project, or the global equivalent) so AI Assistant can find the local server — this file only contains a local command/path, no project data.
Questions about this policy: [email protected] — or see the vendor site linked from the plugin's Marketplace listing (cheburnexus.tech).