Flameox — Claude Code MCP server: what it does, how to install
Coordinates local profiling, benchmark runs, and trace processing for coding agents using direct commands and precise file paths.
claude mcp add flameox -- uvx flameoxWhat Flameox does
The server exposes profiling tools, benchmark targets, and trace parsers directly to coding assistants. It avoids project configuration files and workspace initialization. For artifact analysis, the agent supplies exact filesystem paths; for live capture, it passes a discrete argv list alongside an absolute working directory. Flameox leaves project files untouched and does not crawl parent folders.
Captured metrics live in temporary session memory by default and disappear when the server process shuts down. For durable records, optional preservation writes data bundles indexed by SHA-256 into FLAMEOX_DATA_DIR or the default storage directory. The package requires Python 3.12 or newer to run.
What it helps with
- collect benchmark timings using explicit command arguments and directories
- inspect runtime trace artifacts via exact filesystem paths
- store preserved execution measurements keyed by SHA-256 hashes
Example prompts for the agent
- Run a benchmark capture via Flameox for
python benchmark.pyin the current directory - Inspect the trace artifact at
/tmp/run-trace.jsonto find execution bottlenecks - Check the preserved performance metrics for
evidence_idand show metadata
How to install Flameox
Platform CLI
claude mcp add flameox -- uvx flameoxuvx runs the command named after the package; if the README names another one, use uvx --from flameox <command>
Platform CLI
codex mcp add flameox -- uvx flameoxuvx runs the command named after the package; if the README names another one, use uvx --from flameox <command>
Config file
{
"mcpServers": {
"flameox": {
"command": "uvx",
"args": [
"flameox"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from flameox <command>
Config file
{
"servers": {
"flameox": {
"type": "stdio",
"command": "uvx",
"args": [
"flameox"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from flameox <command>
Platform CLI
gemini mcp add flameox uvx flameoxuvx runs the command named after the package; if the README names another one, use uvx --from flameox <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Flameox
How to install Flameox in Claude Code?
Run claude mcp add flameox -- uvx flameox in your terminal.
What does Flameox do?
It coordinates local profilers and benchmark targets, delivering execution traces and runtime evidence back to the agent.
What are the runtime requirements for Flameox?
It needs Python 3.12 or newer, and target commands must be provided as explicit argv arrays rather than shell strings.
