Pomera AI Commander (PAC)

A desktop text "workbench" + MCP server: clean, transform, extract, and analyze text fast—manually in a GUI or programmatically from AI assistants (Cursor / Claude Desktop / MCP clients).

Stop pasting text into 10 random websites. Pomera (GUI + MCP) - do web searches with MCP and save your work as Pomera Notes in case of text corruption in IDE! Your search API keys are stored encrypted in local database instead of JSON config file.

📊 Why AI needs Pomera! - Pomera's MCP tools reduce token usage upto 70-80% for deterministic operations.

Download latest release · Docs: Tools · MCP Guide · CrewAI Integration · Troubleshooting

60-second demo (what to expect)

Best-for workflows

• Cleaning pasted logs / PDFs (whitespace, wrapping, stats)
• Extracting emails/URLs/IDs via regex
• Normalizing case, sorting, columns
• Hashing/encoding utilities
• Letting Cursor/Claude call these as MCP tools in a repeatable pipeline

MCP Server for AI Assistants

Pomera exposes 22 text processing tools via MCP. Configure your AI assistant:

Cursor (.cursor/mcp.json):

{
  "mcpServers": {
    "pomera": {
      "command": "pomera-ai-commander",
      "timeout": 3600
    }
  }
}

Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "pomera": {
      "command": "pomera-ai-commander",
      "timeout": 3600
    }
  }
}

💡 Tip: If the simple command doesn't work, use the full path. Find it with:

Copy

# For npm install:
npm root -g
# Then use: <result>/pomera-ai-commander/pomera_mcp_server.py

# For pip install:
pip show pomera-ai-commander | grep Location

⏱️ Timeout: The "timeout": 3600 setting (in seconds) prevents MCP request timeouts during long-running AI operations like research and deepreasoning. Cline, Cursor, and Claude Desktop all default to a 60-second timeout, which is too short for AI calls involving web search + deep reasoning (60-300s). See Cline #1306.

See the full MCP Server Guide for Antigravity, executable configs, and troubleshooting.

Prerequisites

Python 3.8+ is required for all installation methods.

macOS (Homebrew)

# Tkinter support (replace @3.14 with your Python version)
brew install python-tk@3.14
pip3 install requests reportlab python-docx

Ubuntu/Debian

sudo apt-get install python3-tk
pip3 install requests reportlab python-docx

Windows

Tkinter is included with Python from python.org.

pip install requests reportlab python-docx

Note: For PEP 668 protected environments, use pip3 install --user or a virtual environment.

Install / Run

Option A — Prebuilt executable (recommended)

Download from Releases and run.

Option B — Python (PyPI)

pip install pomera-ai-commander
# then run:
pomera-ai-commander --help

Option C — Node.js (npm)

npm install -g pomera-ai-commander
# then run:
pomera-mcp --help

Create Desktop Shortcut

After installing via pip or npm, create a desktop shortcut for quick access:

# For pip install:
pomera-create-shortcut

# For npm install (from package directory):
python create_shortcut.py

License

MIT License - see LICENSE for details.