FAQ
Frequently Asked Questions
Everything you need to know about OpenClaw Launch — pricing, features, deployment, and more.
What is OpenClaw Launch?
OpenClaw Launch is a managed hosting platform for open-source AI agents. It runs two frameworks — OpenClaw (MIT license, 270,000+ GitHub stars) and Hermes Agent by Nous Research — and you pick one at deploy time. You configure your agent visually in the browser, click Deploy, and it goes live on managed cloud infrastructure in under 30 seconds, reachable on Telegram, Discord, WhatsApp, and 9+ other channels. There is no Docker to install, no VPS to rent, no command line, and no config files. Each agent runs in its own isolated container with automatic SSL, real-time logs, a web terminal, a file manager, and a live browser view, so you can watch what your agent is doing and take over at any time.
How much does it cost?
There are three tiers. The Free Trial is $0 for 30 minutes with no credit card — web chat, free models, a built-in browser, and full command execution. The Lite plan is $3 for the first month, then $6/month (or $60/year, which works out to $5/month), and includes 1 vCPU, 2 GB RAM, 10 GB storage, one instance, and $1/month in AI credits. The Pro plan is $20/month (or $200/year, $16.67/month effective) with 2 vCPU, 4 GB RAM, 40 GB storage, up to 3 instances, and $10/month in AI credits. Both paid plans include Telegram, Discord, and WhatsApp bots, the Web UI, real-time monitoring, automatic SSL, and thousands of ClawHub skills. Annual plans come with a 30-day refund window. See full pricing.
What AI models are supported?
OpenClaw Launch supports the leading AI models: GPT-5.2 (OpenAI), Claude Opus 4.6 and Claude Sonnet 4.6 (Anthropic), Gemini 3 Pro and Gemini 3 Flash (Google), DeepSeek V4 Pro and V4 Flash, Kimi K2.6, plus a free-tier model for trials. Models route through OpenRouter, and you can switch models at any time from the dashboard without redeploying your agent — the change applies live. If you already pay for ChatGPT Plus or a Claude subscription, you can connect it and use those models at no extra API cost, or bring your own OpenRouter API key for anything else in the OpenRouter catalog. See the model comparison.
Do I need technical knowledge?
No. OpenClaw Launch is built so that non-technical users can run the same open-source agents that normally require DevOps skills. The visual configurator walks you through three choices — framework (OpenClaw or Hermes), AI model, and chat platform — then one click deploys everything. Docker, servers, networking, SSL certificates, monitoring, and framework updates are all handled automatically behind the scenes. Self-hosting the same software typically means renting a VPS, installing Docker, editing JSON config files, and managing SSH and firewalls — roughly 2–6 hours of setup plus ongoing maintenance. On OpenClaw Launch none of that is exposed to you, though power users still get a full web terminal and file manager for direct access when they want it.
How long does deployment take?
About 30 seconds from clicking Deploy to a live agent. The platform provisions your isolated container, applies your configuration, sets up SSL/HTTPS automatically, connects your chosen chat platform, and starts real-time monitoring — all in one automated sequence. After deploying you are redirected to your instances dashboard, where you can open the web chat and send your first message immediately, view live logs, or connect more platforms (Telegram, Discord, WhatsApp, and others) to the same instance. There is no DNS to configure and no certificate wait: the agent is reachable the moment the dashboard shows it running.
Can I use my own API key?
Yes. You can bring your own OpenRouter API key (BYOK) and your agent will bill usage to your own OpenRouter account instead of the included plan credits — useful if you already have OpenRouter credits or need higher usage limits. You can also connect an existing ChatGPT Plus or Claude subscription and use GPT or Claude models at zero extra API cost, since usage rides on the subscription you already pay for. Included plan credits ($1/month on Lite, $10/month on Pro) route through OpenRouter at standard rates with no markup, so you can start without any key and add one later — switching between included credits and BYOK does not require a redeploy.
Is my data secure?
Yes. Every agent runs in its own isolated Docker container with dedicated storage, a separate network namespace, and enforced CPU/memory/disk quotas — containers cannot read each other's data. Configurations and API keys are encrypted with AES-256-GCM; credentials-authenticated users can enable end-to-end encryption where keys are derived from your password (PBKDF2, 100,000 iterations), meaning the platform cannot read your configuration at all. Gateway tokens are cryptographically secure random UUIDs, all traffic uses TLS 1.2+ with HSTS, two-factor authentication is available, and conversations are never used for AI training. Infrastructure runs on Hetzner cloud servers with daily database backups. See our security page.
What chat platforms are supported?
Agents deploy to Telegram, Discord, WhatsApp, WeChat, Slack, and a built-in Web UI, among 9+ supported channels. You can connect multiple platforms to the same instance — one agent answering you on Telegram and your team on Slack, for example — and add or remove platforms after deployment without redeploying. Telegram uses secure pairing, so your bot only responds to you until you explicitly approve other users. Discord bots are invite-only by design. The built-in Web UI works immediately with zero setup and includes file uploads and a live view of what your agent is doing.
What are skills?
Skills are plugins that extend what your AI agent can do. OpenClaw has thousands of community-built skills available on ClawHub, the skills marketplace — covering web search, file management, email, GitHub integration, smart home control, price monitoring, flight check-ins, content creation, and much more. On OpenClaw Launch, skills are browsable and installable from the visual configurator with one click — no command line or code required — and your agent can start using a new skill immediately. Hermes Agent additionally creates and improves its own skills from experience via its self-improving loop, so its capabilities grow with use. Browse all skills.
Can I self-host OpenClaw instead?
Yes. OpenClaw is open-source software under the MIT license, and Hermes Agent is open source too — you can self-host either on your own server with Docker. Expect to provision a VPS (roughly $5–20/month) or always-on home hardware, install Docker, edit JSON configuration, and manage SSL, updates, and uptime yourself; initial setup typically takes 2–6 hours. OpenClaw Launch exists for people who want the same agents without that DevOps work: identical open-source software, but deployed in 30 seconds on managed infrastructure. If you are already self-hosting, you can switch by recreating your configuration in the visual editor — no migration scripts needed. See the managed vs self-hosted comparison.
How do I cancel my subscription?
You can cancel anytime from the Stripe customer portal, accessible from your dashboard settings page — there is no cancellation process to go through and no early-termination fee. Your instance stays active until the end of the billing period you already paid for, so you lose nothing by cancelling early in a cycle. Annual plans additionally include a 30-day refund: if you cancel within the first 30 days, the first month is charged at the standard monthly rate ($6 for Lite, $20 for Pro) and the remaining annual payment is refunded in full.
Do you offer a free trial?
Yes — two ways to try before committing. The Free Trial gives you a real agent for 30 minutes at $0 with no credit card required: web chat, free tokens on capable free models, a built-in browser, and full command execution, so you can test both OpenClaw and Hermes side by side. Premium models (GPT, Claude) and vision are reserved for paid plans. After the trial, the Lite plan at $3 for the first month works as a low-risk extended trial — if it is not for you, cancel before the month ends and you have only paid $3. Annual plans also carry a 30-day refund window.
What is Hermes Agent, and can I host it here?
Hermes Agent is the open-source AI agent framework by Nous Research, and yes — OpenClaw Launch offers fully managed Hermes hosting on the same plans and infrastructure as OpenClaw. Hermes is built for autonomy: persistent multi-level memory that survives across sessions, 40+ built-in tools, and a self-improving loop that creates and refines skills from experience. That makes it a strong choice for long-running agents that need to remember context over weeks of use. You choose between OpenClaw and Hermes with one click at deploy time, and both deploy in under 30 seconds with the same channels, SSL, and monitoring included. See Hermes Agent hosting and the OpenClaw vs Hermes comparison.