OpenClaw — formerly known as ClawdBot, then briefly Moltbot — went from a side project to 147,000 GitHub stars in three months. It's an open-source AI assistant that runs on your own machine, connects to WhatsApp, Telegram, and Discord, and can actually execute tasks on your computer. If you're a business owner wondering whether this thing can replace your customer support team or automate your operations — here's the honest answer.
What is OpenClaw (ClawdBot)?
OpenClaw is a self-hosted AI assistant. You install it on your laptop, server, or cloud VM. It connects to an AI model (Claude, GPT-4, or a free local model via Ollama) and to your messaging platforms. Then you can talk to it through WhatsApp, Telegram, Discord, Slack, Signal, or 15+ other channels.
What makes it different from ChatGPT or other chatbots: it can actually do things on your system. Run shell commands, manage files, control a web browser, schedule tasks, monitor systems. It's not just answering questions — it's executing actions.
The project was created by Austrian developer Peter Steinberger and is fully open-source under the MIT license. It's free. You only pay for the AI model API you choose to connect.
The name changes, explained
- November 2025 — Launched as ClawdBot (a play on Claude + bot)
- January 27, 2026 — Renamed to Moltbot after Anthropic trademark concerns
- January 30, 2026 — Renamed again to OpenClaw (lobster theme continues)
- February 14, 2026 — Steinberger announces he's joining OpenAI; project moves to open-source foundation
How to set up OpenClaw (step by step)
The setup takes about 30 minutes. Here's what you need:
Requirements
- macOS 11+, Linux (Ubuntu 20.04+), or Windows 10+ with WSL2
- 2GB RAM minimum (4GB recommended, 8GB+ if running local AI models)
- 500MB disk space plus model cache
- Node.js 22+
- Stable internet (unless using local Ollama models)
Step 1: Install
On macOS, download the DMG from GitHub releases. On Linux or WSL:
npm install -g clawbot@latest && clawbot initOr use Docker:
docker run -d --name clawbot -v ~/.clawbot:/root/.clawbot -p 18789:18789 ghcr.io/steipete/clawbot:latestStep 2: Start the Gateway
The Gateway is the central nervous system — it routes messages between your channels, AI model, and your system.
clawbot gateway start && clawbot gateway statusStep 3: Connect an AI model
Edit ~/.clawbot/clawbot.json and choose your provider:
| Provider | Cost | Best for |
|---|---|---|
| Anthropic Claude | ~$20–40/mo | Best reasoning, recommended |
| OpenAI GPT-4 | ~$25–50/mo | Versatile, broad model selection |
| Ollama (local) | Free | Maximum privacy, needs 8GB+ RAM |
Step 4: Connect a messaging channel
Telegram is easiest for testing. WhatsApp works but has Terms of Service limitations for automated bots (use a secondary number). Discord requires a bot token from the Developer Portal.
clawbot channel add telegramStep 5: Install skills
OpenClaw has 565+ community-built skills available through ClawdHub. Browse them with clawbot skills browse. Popular ones include Google Calendar integration, GitHub assistant, home automation, and daily briefings.
Real business use cases for OpenClaw
Here's where OpenClaw genuinely shines for small businesses and solopreneurs:
Scheduling and calendar management
Connect to Google Calendar. Ask your AI to schedule meetings, check availability, send reminders — all through WhatsApp or Telegram.
File and document handling
Organize files, rename in bulk, convert formats, extract text from PDFs. Useful for anyone drowning in paperwork.
Social media monitoring
Set up skills to track mentions, summarize trending topics, or draft responses.
Developer workflow automation
GitHub integration, deployment scripts, log monitoring, automated code reviews. This is where OpenClaw excels.
Personal CRM
Track conversations, set follow-up reminders, log client interactions through your messaging app.
System monitoring
Monitor server health, get alerts on downtime, auto-restart services. Great for small DevOps teams.
Where OpenClaw falls short for businesses
OpenClaw is genuinely impressive for personal use and developer workflows. But if you're running a business with real clients, employees, and compliance requirements, there are gaps:
No multi-user access control
OpenClaw is single-user. There's no admin panel, no role-based permissions, no way to control which employees can access which tools. For a team, this is a dealbreaker.
WhatsApp Business API not supported
OpenClaw uses the personal WhatsApp linked devices feature. For business WhatsApp automation — verified sender, message templates, broadcast lists — you need the official Business API with an approved provider.
No enterprise integrations out of the box
Connecting to Salesforce, SAP, QuickBooks, Xero, or your bank's API requires custom development. OpenClaw's skill marketplace focuses on developer tools and personal productivity, not business systems.
No compliance or audit trail
If you handle financial data, healthcare records, or client information, you need encryption at rest, audit logging, data retention policies, and access controls. OpenClaw doesn't provide these.
You maintain everything yourself
Updates, backups, security patches, uptime monitoring — it's all on you. If it breaks at 2am and a client needs something, you're the support team.
No onboarding or training
Setting up OpenClaw requires command-line comfort. You need to configure JSON files, manage API keys, troubleshoot Docker containers. Your operations manager or sales team isn't going to do this.
OpenClaw vs. custom-built AI: when to use which
| Scenario | OpenClaw | Custom AI (e.g. Nora) |
|---|---|---|
| Solo founder or developer | Perfect | Overkill |
| Personal productivity automation | Perfect | Overkill |
| Team of 5+ needs shared AI | No multi-user | Built for this |
| Client-facing WhatsApp bot | TOS risk | Official Business API |
| Bank/invoice/ERP integration | DIY development | Pre-built connectors |
| Compliance requirements | Not available | Role-based access + audit |
| 24/7 uptime required | Self-managed | Managed infrastructure |
| Budget | Free + API costs | $5K–50K project |
The honest take: if you're technical and want a personal AI assistant, OpenClaw is incredible. If you need an AI system that your entire team uses, connects to your business tools, and handles client data securely — you need something custom-built.
Deploying OpenClaw to a server (24/7 availability)
Running OpenClaw on your laptop means it stops when you close your lid. For always-on availability, deploy to a cloud server:
- DigitalOcean Droplet — $6/month for 1GB RAM (enough for Claude/GPT-4 API mode)
- Hetzner Cloud — from €3.79/month, great value in Europe
- AWS EC2 or Lightsail — familiar if you're already in the AWS ecosystem
- Any Linux VPS — install via Docker, use Tailscale or SSH for remote access
For local AI models (Ollama), you need more power — at least 8GB RAM and ideally a GPU. This is where costs go up quickly and cloud hosting becomes less practical.
The ClawdHub skills marketplace
One of OpenClaw's strongest features is the community skill ecosystem. With 565+ skills available, you can extend functionality without writing code. Some highlights:
Need more than OpenClaw can offer?
We build custom AI platforms for businesses — private deployment, enterprise integrations, multi-user access, WhatsApp Business API, and full support. Same philosophy as OpenClaw (your data, your server), but built for teams and real business workflows.
What's next for OpenClaw?
With Steinberger joining OpenAI and the project moving to an open-source foundation, the future is uncertain but promising. The community is massive (147K+ stars), the skill ecosystem is growing, and the foundation structure should ensure long-term sustainability.
For businesses watching from the sidelines: this is worth experimenting with now. Set it up, test it for internal workflows, and understand what AI automation can do for your operations. If you hit the ceiling of what OpenClaw can do alone — that's when a custom solution starts making sense.