JProfiler
# JProfiler MCP Server
[](https://glama.ai/mcp/servers/ej-technologies/jprofiler-mcp)
[JProfiler](https://www.ej-technologies.com/jprofiler) is a Java profiler with AI agent
integration via the [Model Context Protocol](https://modelcontextprotocol.io).
Profile CPU hotspots, memory allocations, JDBC queries, HTTP calls and more
directly from your AI coding assistant.
## Installation
### Claude Code
```
claude mcp add jprofiler -- npx -y @ej-technologies/jprofiler-mcp
```
### Codex
```
codex mcp add jprofiler -- npx -y @ej-technologies/jprofiler-mcp
```
### Cursor
Listed on [cursor.directory](https://cursor.directory/plugins/jprofiler).
[](https://cursor.directory/plugins/jprofiler)
### Other MCP clients
Use the following npx invocation as the MCP server command in your client's configuration:
```
npx -y @ej-technologies/jprofiler-mcp@latest
```
For clients that use a JSON configuration file (Claude Desktop, Antigravity, Windsurf, VS Code, etc.):
```json
{
"mcpServers": {
"jprofiler": {
"command": "npx",
"args": ["-y", "@ej-technologies/jprofiler-mcp"]
}
}
}
```
## License
JProfiler requires a license. On first use, you can start a free 10-day evaluation
or enter an existing license key. Purchase at https://www.ej-technologies.com/store/jprofiler.
License terms: https://www.ej-technologies.com/jprofiler/license
TDQS
Scored across 8 tools
Most tools target distinct stages of the profiling workflow: discovery, session start, attachment, status, heap dump, snapshot loading, and querying. There is some overlap between check_status and attached_action around controlling recording, and between create_heap_dump and attached_action's heap-dump actions, but the descriptions make the intended boundaries reasonably clear.
The set is mostly snake_case, with verb_noun names such as prepare_profiling, load_snapshot, check_status, create_heap_dump, and list_jvms. However, query and attach are bare verbs, and attached_action is an adjective_noun form rather than an imperative verb, so the convention is readable but not fully consistent.
Eight tools is well scoped for a JVM profiling server: each covers a distinct part of the lifecycle from discovering JVMs to recording, loading snapshots, and querying results. There are no obviously redundant tools or missing slots that would suggest the count is off.
The surface covers the full expected workflow: list_jvms for discovery, prepare_profiling and attach for starting sessions, check_status and attached_action for control, create_heap_dump for heap capture, load_snapshot for existing files, and query for analysis. No major lifecycle operation appears to be missing.