What is Clawputer?
Clawputer is a persistent AI assistant that runs continuously inside an OpenComputer cloud sandbox. It routes queries to Claude through OpenRouter, integrates with gbrain for long-term memory, and exposes a Telegram bot interface so you can chat with your always-on assistant from anywhere on any device.
Top Features:
- Always-on sandbox: runs in the cloud 24/7 with no local install or laptop required.
- Claude routing: uses OpenRouter to call Claude models for reasoning, drafting, and tool calls.
- Memory layer: integrates with gbrain to remember past chats, files, and preferences.
- Telegram interface: chat with your assistant from any device using a private Telegram bot.
- Automation hooks: trigger background tasks, scheduled jobs, and indexing of personal knowledge.
Use Cases:
- Personal assistant: ask Claude on the go through a Telegram chat that remembers context.
- Knowledge indexing: feed documents to gbrain and query them later from any chat session.
- Background research: kick off jobs that run while you sleep and report results in Telegram.
- Long-running tasks: run agentic workflows that survive restarts thanks to persistent storage.
- Code automation: wire Claude into scheduled scripts, cron jobs, and backend workflows.
Who Can Use Clawputer?
- Developers: engineers who want a persistent Claude-powered agent running in the cloud.
- Indie makers: solo builders who need an always-on assistant for product and admin tasks.
- Power users: people who chat with Claude all day and want memory across sessions.
- Researchers: teams running long, multi-step agent jobs that need durable execution.
Pricing
- Cloud sandbox (paid): hosted on OpenComputer infrastructure with usage-based cloud sandbox fees.
- OpenRouter usage: pay per Claude token used, charged through your OpenRouter account directly.
- BYO API keys: you bring keys for Claude, OpenRouter, and gbrain, then pay each provider directly.
Pros and Cons
Pros:
- Persistent runtime: the assistant lives in the cloud and survives restarts and disconnections.
- Telegram-first UX: chat with your AI from any device through a familiar messaging interface.
- Memory built-in: gbrain integration gives the agent long-term recall across conversations.
Cons:
- Technical setup: connecting OpenRouter, gbrain, and a Telegram bot requires developer skills.
- Multiple providers: you pay OpenComputer, OpenRouter, and gbrain separately for usage.
- Limited UI: Telegram is the main interface, so visual tools are minimal compared to a web app.
FAQs:
1) What model does Clawputer use?
It routes queries to Claude through OpenRouter, so you get Claude responses by default.
2) Where does my data live?
Memory and knowledge are stored in gbrain while session state runs inside the OpenComputer sandbox.
3) Do I need a Telegram account?
Yes, the main chat interface is a private Telegram bot tied to your account.
4) Is Clawputer open source?
Yes, the Clawputer recipe is shared by OpenComputer for users to inspect and modify freely.
5) Can it run scheduled jobs?
Yes, the sandbox supports background tasks and cron jobs that run even when you are offline.