Forem: 橋本透

WinClaw's 18 Business Plugins: From Invoice Processing to Inventory Management — All AI-Powered

橋本透 — Mon, 02 Mar 2026 11:19:56 +0000

Beyond Development: AI-Powered Business Automation

WinClaw isn't just for developers. It ships with 18 business plugins that automate everyday office tasks using AI. Each plugin works through natural conversation — just describe what you need, and the AI handles the rest.

The 18 Plugins at a Glance

Document & Data Processing

Plugin	What It Does
invoice-processor	Extract data from invoices (PDF/image), validate amounts, export to CSV/Excel
receipt-scanner	Scan receipts, categorize expenses, generate expense reports
document-converter	Convert between formats (PDF, Word, Excel, CSV, JSON) with layout preservation
data-entry-automator	Auto-fill forms and spreadsheets from source documents

Communication & Scheduling

Plugin	What It Does
email-drafter	Draft professional emails from brief notes, with tone/style control
meeting-scheduler	Find optimal meeting times, send invites, handle timezone conversion
minutes-generator	Generate meeting minutes from audio transcripts or notes
translation-bridge	Real-time document translation (JP/EN/ZH) with domain-specific terminology

Finance & Accounting

Plugin	What It Does
budget-tracker	Track project budgets, flag overruns, generate variance reports
payroll-calculator	Calculate wages, deductions, overtime with locale-specific tax rules
expense-reporter	Compile and categorize expenses, generate approval-ready reports

Inventory & Operations

Plugin	What It Does
inventory-manager	Track stock levels, predict reorder points, generate purchase orders
shipping-tracker	Monitor shipments across carriers, alert on delays, update ETAs
order-processor	Process incoming orders, validate stock, generate fulfillment tasks

Reporting & Analytics

Plugin	What It Does
report-generator	Create formatted business reports from raw data with charts
kpi-dashboard	Build real-time KPI dashboards with drill-down capabilities
trend-analyzer	Identify trends in sales/operations data with forecasting
compliance-checker	Verify documents against regulatory requirements (GDPR, SOX, etc.)

How Plugins Work

Each plugin follows a simple pattern:

# Example: Process invoices
winclaw plugin invoice-processor --input ./invoices/ --output ./processed/

# Example: Generate meeting minutes
winclaw plugin minutes-generator --transcript meeting-2024-03.txt

# Example: Track inventory
winclaw plugin inventory-manager --action reorder-check

The AI reads your input, understands the context, processes the data, and outputs structured results.

Real-World Example: Invoice Processing

Drop invoice PDFs into a folder
Run the invoice-processor plugin
AI extracts: vendor, date, line items, amounts, tax
Outputs a clean CSV ready for your accounting software
Flags any discrepancies or missing information

Getting Started

All 18 plugins are included free with WinClaw:

set ANTHROPIC_BASE_URL=https://api.z.ai/api/anthropic
set ANTHROPIC_AUTH_TOKEN=your-api-key
set ANTHROPIC_MODEL=glm-5

Conclusion

WinClaw's 18 business plugins bring AI automation to everyday office tasks. From invoice processing to inventory management, each plugin saves hours of manual work. Best of all, they're completely free and open-source. Give them a try!

WinClaw AI Testing: Fully Automated Web App Testing Without Writing Test Scripts

橋本透 — Mon, 02 Mar 2026 11:10:24 +0000

The Problem with Traditional Testing

Writing Selenium scripts, maintaining Playwright selectors, updating tests every time the UI changes... Manual test maintenance is a massive time sink for development teams.

What if an AI could look at your web app and test it automatically — no scripts needed?

WinClaw AI Dev System Testing

WinClaw is a free, open-source AI development tool for Windows. Its ai-dev-system-testing skill uses Claude Code + Chrome automation to test web applications completely automatically.

How It Works

The testing workflow runs in 5 phases:

Phase 1: Environment Check — Verifies Chrome debug port, MCP connections, and target URLs
Phase 2: Discovery — AI crawls your app, maps all pages, identifies forms and interactive elements
Phase 3: Frontend Testing — Tests UI rendering, responsive design, form validation, navigation
Phase 4: API Testing — Discovers and tests backend endpoints (supports Swagger/OpenAPI auto-detection)
Phase 5: Iteration — Re-tests failed items with different strategies, auto-fixes where possible

What Gets Tested Automatically

Page rendering: Screenshots + accessibility checks on every page
Form validation: Fills forms with valid/invalid data, checks error messages
Navigation: Tests all links, breadcrumbs, routing
Responsive design: Mobile/tablet/desktop viewport testing
API endpoints: GET/POST/PUT/DELETE with schema validation
Error handling: 404 pages, network errors, edge cases

Zero Configuration

Just point WinClaw at your running app:

# Frontend + Backend
winclaw test --frontend http://localhost:3000 --backend http://localhost:3001/docs

# Frontend only
winclaw test --frontend http://localhost:3000

The AI figures out what to test, generates test scenarios, executes them, and produces a detailed report — all without a single line of test code.

Test Reports

After testing completes, you get:

Pass rate by category (UI, API, forms, navigation)
Screenshots of every tested page
Detailed failure logs with reproduction steps
Recommendations for fixes

Getting Started

WinClaw is completely free:

LLM backend (GLM-5, free):

set ANTHROPIC_BASE_URL=https://api.z.ai/api/anthropic
set ANTHROPIC_AUTH_TOKEN=your-api-key
set ANTHROPIC_MODEL=glm-5

Conclusion

WinClaw's AI testing eliminates the need for manual test scripts. Point it at your app, let the AI discover and test everything, and get a comprehensive report. It's like having a QA engineer that never sleeps and never needs training. Try it out!

WinClaw: Auto-Deploy to AWS, Azure & Alibaba Cloud with AI-Powered Conversations

橋本透 — Mon, 02 Mar 2026 11:04:02 +0000

Introduction

"I want to deploy to AWS, but writing CloudFormation YAML is a pain..." "Azure has too many configuration options..." Sound familiar?

I had the same frustrations until I tried WinClaw's cloud auto-deploy skills. Just by having a conversation with AI, I got my app deployed to AWS, Azure, and Alibaba Cloud — fully automated.

What is WinClaw Cloud Deploy?

WinClaw is a free, open-source AI development tool with three cloud deployment skills:

aws-cloud-deploy: Auto-generates CloudFormation templates
azure-cloud-deploy: Auto-generates ARM templates
aliyun-cloud-deploy: Auto-generates ROS templates

All three providers share the same conversational workflow — great for multi-cloud environments.

The Interactive Deploy Experience

Phase 1: Requirements Hearing

The AI asks you natural questions:

Monthly budget: "How much can you spend per month?" → e.g., $50/mo
Expected traffic: "Daily visitors?" → e.g., 1,000 PV/day
Security needs: "Handling personal data?" → Yes/No
Availability: "Acceptable downtime?" → e.g., < 30 min/month

Phase 2: Architecture Proposal

Based on your answers, the AI proposes the optimal architecture from 6 patterns:

Pattern	Cost/mo	Use Case
Lite	$10-40	Single EC2, blogs/testing
Standard	$50-150	EC2 + RDS, small web apps
HA	$150-300	ALB + Auto Scaling + Multi-AZ
Elastic	$250-600	CloudFront + ElastiCache
Serverless	$0-100	Lambda + API Gateway
Container	$300+	ECS/EKS microservices

Security Tiers

Security is also auto-selected from 3 tiers:

Basic (Free): Security Groups + IAM Roles
Standard ($10-30/mo): WAF + VPC Flow Logs + Encryption
Enterprise ($200-3,500/mo): Shield Advanced + GuardDuty + Config Rules + Audit Logs

Phase 3: Automated Deployment

After confirming the architecture, everything runs automatically:

Phase 3A — IaC template generation (CloudFormation/ARM/ROS)
Phase 3B — Code deployment
Phase 3C — Health checks + auto-fix on errors
Phase 3D — Report generation (access URL, cost estimate)

Same Experience on Azure & Alibaba Cloud

Azure generates ARM templates and selects from App Service / VM / AKS / Functions.

Alibaba Cloud generates ROS templates and selects from ECS / FC / ACK. Supports China's MLPS 2.0 compliance.

Getting Started

WinClaw is completely free and open-source:

Use GLM-5 (free) as the LLM backend:

set ANTHROPIC_BASE_URL=https://api.z.ai/api/anthropic
set ANTHROPIC_AUTH_TOKEN=your-api-key
set ANTHROPIC_MODEL=glm-5

Conclusion

WinClaw's cloud deploy skills dramatically lower the barrier to infrastructure setup. Just answer questions about budget and traffic, get an optimal architecture proposal, and watch as everything deploys automatically. Once you try it, there's no going back. Give it a shot!

OpenClaw: An Open-Source AI Assistant That Runs Natively on Windows PC — With Office, Outlook & Multi-Channel Support

橋本透 — Wed, 11 Feb 2026 11:22:35 +0000

TL;DR

OpenClaw (WinClaw) is an open-source personal AI assistant / multi-channel AI gateway that connects to WhatsApp, Telegram, Discord, Signal, Slack, iMessage, Microsoft Teams, and more.

The biggest highlight: it runs natively on Windows PC with a bundled EXE installer (Node.js 22 included, zero prerequisites). It also supports macOS, Linux, and Raspberry Pi.

GitHub: https://github.com/itc-ou-shigou/winclaw
npm: npm install -g openclaw
Docs: https://docs.openclaw.ai
License: MIT

Why OpenClaw?

Every time you want to ask AI something, you open ChatGPT's web interface. What if your AI assistant lived inside the messaging apps you already use every day?

With OpenClaw, you just talk to your AI through WhatsApp, Discord, or Telegram. It responds using Claude, GPT, Gemini, or even local LLMs via Ollama. No browser tabs needed.

Native Windows PC Support

Unlike most AI tools that focus on cloud or Linux, OpenClaw is built with Windows-first thinking:

EXE Installer (Recommended)

Bundled Node.js 22 runtime, zero prerequisites
Download from GitHub Releases
Auto-start via Windows Task Scheduler on login
No admin rights required

Other Install Methods

# PowerShell one-liner
irm https://raw.githubusercontent.com/itc-ou-shigou/winclaw/main/install.ps1 | iex

# winget
winget install WinClaw.WinClaw

# npm
npm install -g openclaw@latest

macOS / Linux

npm install -g openclaw@latest
openclaw onboard --install-daemon

Platform	Daemon Method
Windows	schtasks (Task Scheduler)
macOS	launchd user service
Linux	systemd user service

Skill System: Office & Outlook Automation

OpenClaw's skill system dramatically extends your AI assistant's capabilities:

Built-in Skills

Windows Office — Create/edit Word, Excel, PowerPoint (python-docx, openpyxl, python-pptx)
Outlook Integration — Send/read/manage emails (COM Automation)
Windows Explorer — File search, bulk rename, compression, clipboard
Windows System — Services, processes, registry, event logs
Browser Automation — Control Chrome pages via extension
Coding Agent — Codex CLI, Claude Code programming support

Custom Skills

Skills are SKILL.md + scripts. Create and share freely. Community skills on ClawHub.

Supported AI Models

Anthropic Claude — Opus 4.6, Sonnet, Haiku (OAuth + API key)
OpenAI — GPT-4o, o1, o3, Codex
Google Gemini — 2.5 Pro, Flash
Local LLMs — via Ollama

Supported Messaging Channels

Multiple channels simultaneously.

Agent Capabilities

File operations, web search, browser automation
Shell command execution
Scheduled tasks (cron)
Camera/screen capture (node pairing)
Long-term memory system
Sub-agent spawning

Quick Start

Windows

winclaw onboard --flow quickstart
winclaw daemon install
winclaw dashboard

macOS / Linux

openclaw onboard --install-daemon
openclaw dashboard

Summary

OpenClaw aims to be the AI assistant that runs on your own PC (especially Windows). With built-in skills for Office and Outlook, it directly supports your daily workflow.

GitHub: https://github.com/itc-ou-shigou/winclaw
npm: https://www.npmjs.com/package/openclaw
Docs: https://docs.openclaw.ai
Discord: https://discord.com/invite/clawd

WinClaw: I Built an Open-Source AI Assistant That Connects to All Your Messaging Apps

橋本透 — Wed, 11 Feb 2026 08:50:41 +0000

TL;DR

I built WinClaw, an open-source personal AI assistant that runs locally on your machine and connects to 11+ messaging channels. Think of it as a self-hosted AI gateway that lets you talk to Claude, ChatGPT, or any AI model through WhatsApp, Telegram, Slack, Discord, Signal, and more.

GitHub: github.com/itc-ou-shigou/winclaw

The Problem

AI assistants are trapped in their own apps. Want to use Claude? Open claude.ai. ChatGPT? Open chatgpt.com. But what if you want your AI assistant available everywhere you already chat — WhatsApp with friends, Slack at work, Discord in communities, Signal for privacy?

The Solution: WinClaw

WinClaw is a personal AI gateway that:

Connects to 11+ channels — WhatsApp, Telegram, Slack, Discord, Signal, iMessage, Microsoft Teams, Matrix, Google Chat, Zalo, and WebChat
Runs on any OS — Windows (standalone EXE installer), macOS (launchd), Linux (systemd)
Supports multiple AI models — Claude, OpenAI, or any OpenAI-compatible provider with automatic failover
Keeps data local — Everything stays on your machine. No cloud dependency.
5-minute setup — Interactive onboarding wizard guides you through everything

Key Features

🧠 Memory System

Your assistant remembers context across sessions with daily notes and long-term memory files.

🔧 Extensible Skills

Browse 100+ skills at clawhub.com or create your own. Skills let your assistant do everything from managing emails to controlling smart home devices.

🌐 Browser Automation

Chrome Browser Relay lets your assistant browse the web, fill forms, and automate tasks in your browser.

📱 Device Pairing

Pair phones, cameras, and IoT devices. Your assistant can snap photos, check screens, and control hardware.

⏰ Proactive Scheduling

Cron jobs and heartbeat monitoring let your assistant check emails, calendars, and notifications without being asked.

🔄 Multi-Model Failover

Configure multiple AI providers with automatic failover. If one goes down, your assistant seamlessly switches.

Installation

# Windows (recommended)
winget install WinClaw.WinClaw

# macOS / Linux
npm install -g openclaw@latest
openclaw onboard --install-daemon

# Docker
docker pull openclaw/openclaw:latest

On Windows, you can also download the standalone EXE installer from GitHub Releases — zero prerequisites, bundled Node.js runtime.

Tech Stack

Runtime: Node.js 22 LTS
Recommended Model: Anthropic Claude Opus 4.6
License: MIT
Platforms: Windows 10+, macOS, Linux
Install methods: EXE, npm, winget, Docker, Nix

What's Next

I'm actively working on:

More skills and integrations
Improved multi-language support
Community skill marketplace
Better documentation and tutorials

Try It Out

👉 GitHub: github.com/itc-ou-shigou/winclaw
📚 Docs: docs.openclaw.ai
🧩 Skills: clawhub.com
💬 Community: Discord

Feedback, issues, and PRs are all welcome! Would love to hear what you think. 🧡