MCP Tool Server

FrankenClaw

Give your AI agent real tools — without losing control of cost or models.

Create your own FrankenClaw Agent using FrankenTools!

FrankenClaw is a controlled execution layer for AI agents. It serves FrankenTools over standard MCP — search, vision, browser automation, web scraping, Shopify, and content generation — so any agent can use them. You pick the models. You control the keys. The tools just execute.

⚠ Important: This project is not affiliated with any cryptocurrency, token, or investment scheme. FrankenClaw is an open-source MCP toolbox for AI agents, built by Project Sparks. If anyone offers you a “FrankenClaw token” — it’s a scam.

View on GitHub Install

Security

Your machine. Your keys. Your rules.

Runs Locally

Executes on your machine, not in the cloud.

No Key Duplication

API keys live in one place, read at runtime.

You Control Providers

Pick which models handle which jobs.

No Hidden Costs

Route cheap to grunt work, smart to decisions.

12 FrankenTools

Hands your agent can actually use.

Each tool is a single Python file. Auto-discovered on startup. No registration, no config.

🔍

search_web

Web search via SearXNG. Private, fast, no API key needed.

👁

vision_analyze

Read any image on disk. Sends to a vision model via OpenRouter.

🌐

browser_task

Full browser automation. Click, fill, navigate, screenshot — in a visible window.

📄

web_scrape

Clean page scraping via Firecrawl. Strips ads and nav, returns markdown.

🛒

shopify_products

List, create, update, delete Shopify products. Full catalog access.

📦

shopify_inventory

Manage inventory levels and locations across warehouses.

🏷

shopify_collections

Browse custom and smart product collections.

🎙

notebooklm_manage

Create, list, rename, and delete Google NotebookLM notebooks.

📎

notebooklm_source

Add URLs, text, or files as sources. List and manage notebook content.

⚡

notebooklm_generate

Generate audio podcasts, videos, infographics, slide decks, and more.

📥

notebooklm_download

Download generated artifacts — MP3, MP4, PNG, PDF, CSV.

💬

notebooklm_ask

Ask questions grounded in your notebook sources. AI-powered answers.

Architecture

Pure function. No opinions.

FrankenClaw does one thing: execute tool calls and return results. Everything else belongs to the agent.

No Memory

The agent has Mnemo Cortex for that.

No Models

Rocky's Switch handles model selection.

No Context

OpenClaw manages conversations.

Auto-Discovery

Drop a .py file in tools/. It just works.

Install

Three commands.

git clone https://github.com/GuyMannDude/frankenclaw.git
cd frankenclaw
pip install -r requirements.txt

openclaw mcp set frankenclaw "python3 /path/to/frankenclaw/server.py"
openclaw gateway restart

API keys come from Rocky's Switch (~/.rockys-switch/keys.json). Tool settings live at ~/.frankenclaw/config.json, auto-created on first run.

Or install the full stack with one command:

curl -fsSL https://projectsparks.ai/install.sh | bash

The Sparks Stack installer sets up FrankenClaw + Rocky's Switch + Mnemo Cortex and wires them together.

FrankenClaw

Your machine. Your keys. Your rules.

Hands your agent can actually use.

Pure function. No opinions.

Three commands.

Part of the family.