Contextpulse MCP Server

Screen capture, voice dictation, clipboard, keyboard/mouse activity. All local, all private.

Local serverstdioPython

What is the Contextpulse MCP server?

Most knowledge and memory work still happens through a UI a human drives. Contextpulse MCP server moves it into the conversation instead. Screen capture, voice dictation, clipboard, keyboard/mouse activity. All local, all private.

The short version

  • Local-first, zero cloud dependency. — Your screen, voice, and input data never leave your machine. No accounts, no subscriptions, no third-party servers. Privacy by architecture, not by policy
  • MCP-native from day one. — ContextPulse exposes all context as MCP tools. Any MCP client (Claude Desktop, Cursor, Windsurf, VS Code) gets full context without custom integrations
  • True multi-modal in a single daemon. — Screen capture, voice dictation, keyboard/mouse input, and semantic memory run in one lightweight process (<1% CPU). No stitching multiple tools together
  • Open source (AGPL-3.0). — Fully auditable, self-hostable, and extensible. No vendor lock-in, no SaaS dependency, no risk of acquisition-driven shutdowns

The tools it exposes

The server publishes 14 tools. What each one is for:

  • get_screenshot — Capture screen (active monitor, all monitors, or a region)
  • get_recent — Recent frames from the rolling buffer (with diff filtering)
  • get_screen_text — OCR the current screen at native resolution
  • get_monitor_summary — Lightweight text summary of all monitors (low token cost)
  • get_buffer_status — Daemon health check + buffer stats
  • get_activity_summary — App usage breakdown over last N hours
  • search_history — Full-text search across window titles + OCR text
  • get_context_at — Frame + metadata from N minutes ago
  • get_clipboard_history — Recent clipboard entries
  • search_clipboard — Search clipboard by text content
  • get_agent_stats — Which MCP clients are consuming context, and how often
  • get_recent_transcriptions — Recent voice dictation history (raw + cleaned)

Getting it running

contextpulse-memory on PyPI is all you need. Most clients run it directly, so configuration is a few lines and a restart.

How it compares

Plenty of knowledge and memory servers cover similar ground. The differences that matter in practice are scope of access and how much setup stands between you and a working tool call. Contextpulse's toolset — get_screenshot, get_recent, get_screen_text and 11 more — is a fair guide to whether it matches your workflow. It is maintained by contextpulse; worth a glance at recent repository activity before you build anything load-bearing on it.

SyncDev reviews every entry in this directory against the project's own documentation before publishing, and revisits them as servers change.

Things to watch

  • It runs with your machine's permissions. That is convenient and also the reason to think about what you point it at before you approve a tool call.
  • With 14 tools registered it takes up a noticeable share of the context window; turn it off in projects that never touch Contextpulse.
  • Keep per-call confirmation enabled while you learn its behaviour; it is the cheapest safeguard you have.

Available tools

ToolWhat it does
get_screenshotCapture screen (active monitor, all monitors, or a region)
get_recentRecent frames from the rolling buffer (with diff filtering)
get_screen_textOCR the current screen at native resolution
get_monitor_summaryLightweight text summary of all monitors (low token cost)
get_buffer_statusDaemon health check + buffer stats
get_activity_summaryApp usage breakdown over last N hours
search_historyFull-text search across window titles + OCR text
get_context_atFrame + metadata from N minutes ago
get_clipboard_historyRecent clipboard entries
search_clipboardSearch clipboard by text content
get_agent_statsWhich MCP clients are consuming context, and how often
get_recent_transcriptionsRecent voice dictation history (raw + cleaned)
get_voice_statsDictation count, duration, accuracy stats
get_vocabularyCurrent word correction entries

How to install the Contextpulse MCP server

{
  "mcpServers": {
    "contextpulse": {
      "type": "http",
      "url": "http://127.0.0.1:8420/mcp"
    }
  }
}

Configuration as documented by the project. Restart the client after saving.

Example prompts to try

  • Use Contextpulse to get screenshot.
  • Use Contextpulse to get recent.
  • Use Contextpulse to get screen text.

Frequently asked questions

It connects Contextpulse to MCP-compatible AI assistants such as Claude and Cursor, exposing 14 tools (get_screenshot, get_recent, get_screen_text, and more) that the assistant can call on your behalf. Instead of copying data back and forth by hand, the assistant works with Contextpulse directly.