mcpdirectory.dev
9,129 servers · updated 23 Sept

Eeg MCP server

official from AImplifier Active last commit 59 d
{
  "mcpServers": {
    "aimplifier-eeg": {
      "command": "uvx",
      "args": [
        "eeg-mcp"
      ]
    }
  }
}
file .mcp.json (in your project root)install uvx eeg-mcpversion 0.1.0

What it does

An MCP server that exposes real-time EEG workflows for AI agents, including acquisition from 66 BrainFlow-compatible boards, wall-clock replay of recordings, online DSP processing, live browser monitoring, recording to standard formats, and controlled stimulation delivery. Designed for BCI researchers, clinicians, and developers building EEG-driven applications.

  • Stream live EEG from hardware boards or replay recordings at their original rate, with the agent querying signal quality and band power before proceeding.
  • Apply online filters and custom DSP processors to the incoming signal stream while recording to crash-safe MNE `.fif` files with event logs attached.

Tools this server exposes

ToolDoes
start_streamBegin acquiring or replaying EEG from a specified board or file.
check_signal_qualityReturn per-electrode signal quality metrics from the current stream.
set_filtersApply bandpass, notch, and other online filters to the acquisition pipeline.
get_band_powerRetrieve power spectral density in standard frequency bands over a time window.
start_monitorLaunch a token-gated live browser view of rolling traces, events, and band power.
start_recordingWrite incoming EEG and events to a `.fif` file with attached metadata.
send_stim_eventDispatch a stimulation event via LSL, TTL, serial, or hardware-specific backend.
add_custom_processorRegister a user-defined real-time DSP or feature-extraction plugin into the acquisition loop.

Before you enable it

An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.