Managed OpenClaw hosting built for real Telegram use

Telegram is fast to set up, stable for long-running bots, and familiar to most users. It gives you a simple path to go live first, then expand to more channels later.

Bot Token is your bot's secret key from @BotFather. User ID is the numeric ID of who is allowed to chat with your bot - in most cases that's your own Telegram User ID (find it via @Getmyid_bot or @userinfobot). Token identifies the bot; User ID controls access.

Relaunch means starting your bot on a fresh server with your latest settings. Use it when a server has issues, disk is full, you want a new bot token, or you want a clean restart. You get up to 10 free relaunches every 30 days, with optional encrypted memory restore.

If you want the fastest openclaw setup and less ops work, cloud is usually better: install openclaw is automated, uptime is easier, and relaunch/recovery is built in. Mac mini is great when you want full local control and already manage your own hardware. In practice, many users start in cloud and move parts in-house later.

If you are unsure, start with the default model that ClawSimple preselects for your plan. It is chosen for balanced quality, speed, and cost. You can switch models anytime later.

Our billing is transparent and real-time: if you use the BYOM plan, we charge for setup and operations. If you use the built-in AI quota, you are charged based on actual OpenRouter model usage costs (we automatically choose the most cost-effective provider); on top of that, we add only a 15% forwarding service fee.

We added BYOK to every plan because OpenClaw can use a lot of tokens in real usage, and many users get better value from their own Codex, Bailian, or similar subscriptions. BYOK means you can connect your own AI account. BYOM means your own account is the main AI source for that plan. Standard and Max users can still stay on the built-in default model if they want, and the current defaults such as Gemini 3.1 Flash can usually support thousands to tens of thousands of interactions before you need to switch.

Yes. Shared is an officially supported plan operated with security best practices. Each user runs in an isolated Docker environment, and we apply resource limits per container to reduce noisy-neighbor impact. There can still be short performance fluctuations when other workloads on the same host are very busy, but the limits and ongoing monitoring are designed to keep that impact minimal.

Not as a separate server subscription. An extra agent reuses the same deployment and server, so you are not buying another server just to add one more bot. What does grow is usage: more agent traffic can consume the deployment's managed AI credits faster, or increase usage on your own AI account if you run BYOM.

At minimum, a fresh Telegram Bot Token for that new bot and the AI choice for that agent. The allowlist usually defaults to your saved Telegram User ID, but you can change it. If the agent uses BYOM, the server-level BYOM connection must already exist and the agent still needs a model selection.

Go to Profile > Overview, open the deployment card, click Add Agent, fill in the bot token and model settings, then submit. ClawSimple queues an add_agent job, updates the running OpenClaw config on that same deployment, and the new bot appears as active once the job finishes.

There is no single fixed number in the current product logic. A few light Telegram assistants on one deployment are usually fine; pressure shows up when several agents are all active at once, especially if they run long tool calls or heavier models. Managed AI credits are tracked at the deployment or seat level, so agents on the same deployment share that usage bucket.