Question 1

What is OpenClaw?

Accepted Answer

OpenClaw is an open-source personal AI assistant with 380K+ GitHub stars. It runs on your own server and connects to messaging platforms like WhatsApp, Telegram, Discord, and Slack. You control the data, the model, and the integrations.

Question 2

What is OpenClawPro?

Accepted Answer

OpenClawPro is a professional installation and management service for OpenClaw. We handle the deployment, security hardening, persistent memory setup, and ongoing support so you can use OpenClaw without touching a terminal.

Question 3

What is the link between OpenClaw and OpenClawPro?

Accepted Answer

OpenClawPro is an independent service provider. We are not affiliated with the official OpenClaw project by Peter Steinberger. We admire the open-source work and make it accessible to non-technical users through professional deployment.

Question 4

Is OpenClawPro officially affiliated with OpenClaw?

Accepted Answer

No. OpenClawPro is operated by Delta-One Capital Sàrl, an independent company. We use the open-source OpenClaw software under its permissive license and provide value-added services: installation, security, memory, and support.

Question 5

Is OpenClaw free?

Accepted Answer

The OpenClaw software itself is free and open-source. You can download and install it yourself at no cost. OpenClawPro charges for professional installation, security auditing, persistent memory configuration, and ongoing support.

Question 6

What is the real monthly cost to run OpenClaw?

Accepted Answer

VPS hosting runs $7-24/month depending on your provider. AI model API costs vary: Haiku 4.5 costs roughly $0.001 per message, Sonnet 5 about $0.01, and Opus 4.8 around $0.05. Most users spend $5-15/month total on hosting plus API calls.

Question 7

Do I need to know how to code?

Accepted Answer

Not with OpenClawPro. We handle every technical step: server provisioning, Docker deployment, security hardening, SSL certificates, memory layers, and channel connections. You receive a working assistant ready to use from day one.

Question 8

Which operating systems does OpenClaw support?

Accepted Answer

OpenClaw runs inside Docker containers, so it works on any Linux VPS (Ubuntu, Debian, CentOS), macOS, and Windows with WSL2. Our installations target Ubuntu 22.04+ on VPS for maximum stability.

Question 9

Is OpenClawPro available in French?

Accepted Answer

Yes. Our website supports 9 languages: English, French, German, Spanish, Italian, Portuguese, Russian, Chinese, and Japanese. Your AI assistant also responds in any language supported by the underlying model, including French.

Question 10

Which AI models are compatible?

Accepted Answer

OpenClaw supports all major providers. Anthropic: Claude Fable 5, Opus 4.8, Haiku 4.5. OpenAI: GPT-5.6 Sol, GPT-5.6 Luna. Google: Gemini 3.1 Pro and Flash. Amazon Nova Premier and Cohere Command A for enterprise stacks. Plus open champions (DeepSeek V4, Qwen 3.6, Llama 4, Mistral Large 3) served at high speed via Groq or NVIDIA NIM, and any model through Ollama for fully local/offline operation.

Question 11

Can I install OpenClaw on a VPS?

Accepted Answer

Yes, and we recommend it. A VPS gives you 24/7 uptime, a dedicated IP, and full control. We install on any Linux VPS from providers like Hostinger, DigitalOcean, Hetzner, or Vultr. The process takes 24-48 hours after purchase.

Question 12

What is the minimum server configuration?

Accepted Answer

2 vCPU and 4 GB RAM handles most workloads comfortably. For heavy multi-channel usage or local model inference, we recommend 4 vCPU and 8 GB RAM. Storage needs are minimal: 20 GB SSD is sufficient.

Question 13

How do I connect WhatsApp to OpenClaw?

Accepted Answer

We configure the WhatsApp Business API bridge during installation. You scan a QR code with your phone, and the connection persists across reboots. Our Advanced and Managed plans include OAuth auto-refresh so the session never expires.

Question 14

How do I connect Telegram?

Accepted Answer

We create a Telegram bot via BotFather and wire it into your OpenClaw instance. The bot token is stored securely in environment variables. Messages flow in real time with no polling delay.

Question 15

What is persistent memory and why does it matter?

Accepted Answer

Without persistent memory, your AI forgets everything after each conversation. OpenClaw's Supermemory system provides long-term recall across sessions, unlike classic approaches that forget after compaction. Your assistant remembers preferences, past conversations, and context.

Question 16

What is context compaction?

Accepted Answer

Context compaction compresses long conversation histories into concise summaries so the AI stays within its token window without losing track of prior exchanges. This keeps responses fast and relevant even after hundreds of messages.

Question 17

What is Supermemory?

Accepted Answer

Supermemory is OpenClaw's 6-layer memory architecture. It stores short-term context, long-term facts, user preferences, conversation summaries, procedural knowledge, and semantic embeddings. The result: persistent, long-term recall, meaning your assistant genuinely remembers you across sessions.

Question 18

Can OpenClaw run fully offline with local models?

Accepted Answer

Yes. Using Ollama, you can run models like Llama 4, Mistral, or Qwen 3 entirely on your own hardware. Zero data leaves your network. You need a capable GPU (8+ GB VRAM) or a modern CPU with 16+ GB RAM for acceptable speed.

Question 19

How are updates handled?

Accepted Answer

On our Managed plan, we apply updates automatically with zero-downtime rolling deployments. On Starter and Advanced plans, we provide a one-command update script. We test every OpenClaw release before recommending it.

Question 20

Docker vs. native installation: which should I choose?

Accepted Answer

Docker. We deploy exclusively via Docker Compose for isolation, reproducibility, and easy rollback. Native installations create dependency conflicts and make updates risky. Docker adds negligible overhead on modern Linux kernels.

Question 21

Is OpenClaw safe to run?

Accepted Answer

It can be, but most installations are not. Security research (Maor Dayan, corroborated by CSO Online) found roughly 42,665 OpenClaw instances exposed on the open internet, with around 93% exhibiting authentication-bypass conditions. A default install is a security risk. Our 12-point audit and managed updates eliminate these vulnerabilities.

Question 22

What is ClawBleed (CVE-2026-25253)?

Accepted Answer

ClawBleed is a real, actively-exploited OpenClaw vulnerability rated CVSS 8.8 (HIGH). Per the NVD, OpenClaw before 2026.1.29 reads a gatewayUrl from a query string and automatically opens a WebSocket connection without prompting, sending a token. The result is a cross-site WebSocket hijack enabling one-click remote code execution and auth-token theft. It is fixed in OpenClaw 2026.1.29 (NVD; CCB Belgium advisory). It should not be confused with ClawHavoc, a separate supply-chain campaign involving malicious skills in the ClawHub registry.

Question 23

Are there really ~42,665 exposed instances?

Accepted Answer

That figure comes from early-2026 security research (Maor Dayan, with corroborating CSO Online scans), which found roughly 42,665 publicly accessible OpenClaw instances, around 93% of them exhibiting authentication-bypass conditions. Separately, the ClawHavoc campaign (Antiy CERT, February 2026) placed 1,184+ malicious skills in the ClawHub registry (roughly 247,000 cumulative installs), several distributing the Atomic macOS Stealer (AMOS).

Question 24

How does OpenClawPro secure my installation?

Accepted Answer

Our 12-point security audit covers: firewall rules, fail2ban, SSH hardening, reverse proxy with rate limiting, TLS/SSL certificates, authentication enforcement, environment variable encryption, Docker network isolation, automatic security updates, ClawHub skill vetting, backup encryption, and intrusion detection logging.

Question 25

What about prompt injection attacks?

Accepted Answer

We configure input sanitization, output filtering, and system-prompt anchoring to mitigate prompt injection. We also disable or vet third-party ClawHub skills that could introduce malicious instructions. No system is 100% immune, but we reduce the attack surface significantly.

Question 26

Are ClawHub skills safe to use?

Accepted Answer

Not by default. The ClawHavoc campaign alone planted 1,184+ malicious skills in the ClawHub marketplace (Antiy CERT, February 2026). We review and whitelist every skill before enabling it on your instance. Unvetted skills can exfiltrate data or override your system prompt.

Question 27

Is OpenClawPro GDPR compliant?

Accepted Answer

We help you deploy OpenClaw on Swiss or EU-based servers (Infomaniak, Hetzner, OVH) if data residency matters to you. All data stays on your VPS. We never access your conversations or personal data. Our website uses consent-gated analytics behind a cookie banner: nothing loads before you accept.

Question 28

Why choose OpenClawPro over a DIY installation?

Accepted Answer

A DIY install takes 4-8 hours if you know Docker, Linux, and networking. Most fail on persistent memory, security, or OAuth refresh. We deliver a production-ready instance in 24-48 hours with 12-point security, Supermemory, and multi-channel setup included.

Question 29

How does OpenClawPro compare to MyClaw?

Accepted Answer

MyClaw charges $29-239/month (public prices as of July 2026, tokens not included) as a hosted SaaS. You share infrastructure and have limited customization. OpenClawPro gives you your own server, full model choice, unlimited messages, and total data ownership. After 8-15 months on a comparable tier, our one-time plans cost less than MyClaw's subscription.

Question 30

Is the payment one-time or recurring?

Accepted Answer

Self-Hosted plans (Starter $169, Advanced $279, Scale $499) are one-time payments. You own the installation forever. Managed plans (Essential $55/mo, Pro $129/mo, Business $219/mo) are recurring because they include hosting, monitoring, updates, and priority support.

Question 31

What does the Starter plan include?

Accepted Answer

Starter gives you a full OpenClaw deployment on your VPS with 1 messaging channel, our 12-point security audit, a basic installation guide, and email support. It covers the essentials for someone who wants a secure, working instance.

Question 32

What does the Advanced plan include?

Accepted Answer

Advanced adds 6-layer persistent memory (Supermemory), multi-channel setup (WhatsApp, Telegram, Discord, Slack), OAuth auto-refresh, 30-day priority support, custom AI persona configuration, and a training session. This is our most popular plan.

Question 33

What does the Managed plan include?

Accepted Answer

Managed includes everything in Advanced plus: VPS hosting included and managed by us, automatic updates with zero-downtime deployment, 24/7 monitoring, priority incident response, and monthly optimization reviews. You never touch a server.

Question 34

Do you offer Enterprise plans?

Accepted Answer

Yes. Enterprise covers multi-tenant deployments, custom SLAs, white-label solutions, dedicated account management, and compliance documentation. Contact us for a tailored quote based on your organization's requirements.

Question 35

How fast is the delivery?

Accepted Answer

Starter and Advanced installations are delivered within 24-48 hours after payment. Managed plans also go live within 24-48 hours of checkout. We provision and configure the hosting infrastructure for you. Enterprise timelines depend on scope.

Question 36

What happens after installation?

Accepted Answer

Starter includes email support. Advanced gives you 30 days of priority support. Managed provides ongoing 24/7 monitoring and support. All plans include our installation documentation and a quick-start guide so you can start using your assistant immediately.

Question 37

Do you accept cryptocurrency payments?

Accepted Answer

Not at the moment: payment is by card through Stripe, settled in your local currency (CHF, EUR, USD and more). We're evaluating crypto payment via Coinbase Commerce; contact us if it's a requirement for you and we'll arrange it.

Question 38

Can OpenClaw handle crypto trading?

Accepted Answer

OpenClaw can connect to exchange APIs (Binance, Kraken, Coinbase) via custom skills and execute trades based on your instructions. We configure the API keys securely and set up safeguards. This requires the Advanced or Managed plan for the persistent memory needed to track positions.

Question 39

Can I run multiple AI agents in parallel?

Accepted Answer

Yes. OpenClaw supports spawning parallel agent threads that work on different tasks simultaneously. One agent researches while another drafts, or one monitors prices while another handles customer messages. Each agent maintains its own context within the shared memory system.

Question 40

How does OpenClaw compare to Claude Code?

Accepted Answer

Claude Code is Anthropic's CLI coding assistant. OpenClaw is a general-purpose AI assistant that connects to messaging platforms, has persistent memory, and supports 191 integrations beyond coding. They serve different purposes: Claude Code excels at software development, OpenClaw excels at being your always-on personal assistant across all channels.

Frequently Asked Questions

General

Technical

Security

Commercial

Advanced

Still have questions?