v1.7.0 — Latest

SDK & CLI Downloads

Integrate Arch Tools into any project — REST SDK, MCP server, or standalone CLI. Runs anywhere AI agents run.

Quickest path to 64 AI tools

Pick your package manager and get started in 30 seconds

npm install @deesmo/arch-tools-mcp
CLI Downloads
Standalone binaries — no Node.js required. Download, chmod, run.
🪟
Windows
Native x64 binary for Windows 10 and 11. Run from PowerShell or CMD.
arch-tools-cli-win-x64.exe · v1.7.0
Download .exe
🍎
macOS
Universal binary for Apple Silicon (M1/M2/M3) and Intel Macs.
arch-tools-cli-macos · v1.7.0
Apple Silicon (M1/M2/M3) Intel (x64)
🐧
Linux
x64 and ARM64 binaries for Debian, Ubuntu, RHEL, Alpine, and more.
arch-tools-cli-linux · v1.7.0
x64 Binary ARM64 (Raspberry Pi / servers)
After downloading on Mac/Linux, make it executable:
chmod +x arch-tools-cli-* && ./arch-tools-cli-* --help
🔌
MCP Server — Model Context Protocol
Give Claude, GPT-4, Cursor, and any MCP-compatible AI instant access to all 64 Arch Tools. One install, 64 tools appear in your AI's context.
STEP 1 — INSTALL
Install the MCP package
npm install -g @deesmo/arch-tools-mcp
STEP 2 — CONFIGURE
Add to your AI client config
arch-tools-mcp --api-key YOUR_KEY
STEP 3 — CLAUDE DESKTOP
claude_desktop_config.json
{
  "mcpServers": {
    "arch-tools": {
      "command": "arch-tools-mcp",
      "env": {
        "ARCH_API_KEY": "your_key"
      }
    }
  }
}
STEP 4 — STREAMABLE HTTP
Use the hosted MCP endpoint
https://arch-tools-mcp.onrender.com/mcp
Pass x-api-key: YOUR_KEY header
Supported AI Clients & Models
Use the MCP server with any AI client, or call the REST API directly from any AI model that supports tool use.

MCP Clients

Claude Desktop
✓ MCP
Cursor
✓ MCP
Windsurf
✓ MCP
Continue.dev
✓ MCP
Zed
✓ MCP
Cline
✓ MCP
OpenClaw
✓ MCP
Any MCP Client
✓ MCP

AI Models (via REST + ai-generate tool)

Claude 3.5 / 4
✓ REST
GPT-4o / o1
✓ REST
Grok (xAI)
✓ REST
Gemini 2.0
✓ REST
REST SDK — Code Examples
Integrate directly via HTTP — no SDK required. Works in any language.
JavaScript / Node.js
# No SDK needed — plain fetch works
const res = await fetch(
  'https://archtools.dev/v1/tools/search-web',
  {
    method: 'POST',
    headers: {
      'x-api-key': 'YOUR_KEY',
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({ query: 'AI agents' })
  }
);
const data = await res.json();
Python
# pip install requests
import requests

r = requests.post(
  'https://archtools.dev/v1/tools/search-web',
  headers={'x-api-key': 'YOUR_KEY'},
  json={'query': 'AI agents'}
)
data = r.json()
Go
# Standard library only
req, _ := http.NewRequest(
  "POST",
  "https://archtools.dev/v1/tools/search-web",
  strings.NewReader(`{"query":"AI"}`),
)
req.Header.Set("x-api-key", "YOUR_KEY")
resp, _ := http.DefaultClient.Do(req)
cURL
# Works from any terminal
curl -X POST \
  https://archtools.dev/v1/tools/search-web \
  -H "x-api-key: YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "AI agents"}'
Copied!