# Gcore Gclaw

> Managed OpenClaw service with built-in inference for launching AI agents instantly.

- [Gclaw](https://docs.gcore.com/gclaw.md): Managed OpenClaw service with built-in inference for launching AI agents instantly.
- [Getting started](https://docs.gcore.com/gclaw/getting-started.md): Create and launch a Gclaw AI assistant instance in the Gcore Customer Portal, access the chat interface, and manage instance settings including Restart, Backup, Restore, and Recreate operations.
- [Manage instance](https://docs.gcore.com/gclaw/manage-instance.md): Manage Gclaw instance restart, backup, restore, and recreation operations via the Settings section in the sidebar; configure data persistence for workspace files, configuration, API keys, skills, and chat history across restart and recreate actions.
- [Configure API keys](https://docs.gcore.com/gclaw/api-keys.md): Configure API keys for Gclaw coding agents and skills by providing credentials directly in the chat interface for secure storage in instance configuration.

## Integrations
- [Connect Telegram](https://docs.gcore.com/gclaw/connect-telegram.md): Configure Telegram bot integration with Gclaw by creating a bot via BotFather, adding the bot token to OpenClaw configuration via the Gclaw chat, and pairing Telegram accounts using pairing codes.
- [Connect Discord](https://docs.gcore.com/gclaw/connect-discord.md): Configure Gclaw Discord integration by creating a Discord bot application in the Developer Portal, enabling Message Content Intent, adding the bot token to OpenClaw configuration via the Gclaw chat, and authorizing the bot to Discord servers.
- [Security model](https://docs.gcore.com/gclaw/security.md): Configure OpenClaw account isolation with dedicated VMs, unique API keys per account, and network isolation from Gcore backend; manage shared security responsibilities including prompt safety, credential protection, third-party bot token security, and agent activity monitoring.
- [Limits and quotas](https://docs.gcore.com/gclaw/limits.md): View Gclaw inference limits including context window, maximum output, reasoning capability, and account-tier-dependent rate limits on DeepSeek V4 Flash model running on Gcore H200 GPUs.
- [Pricing](https://docs.gcore.com/gclaw/pricing.md): Gclaw offers Lite ($10/€9) and Standard ($30/€26) monthly plans with a 7-day trial; both include managed inference, Telegram and Discord integrations, and Clawhub skills; overage at $0.11/€0.10 per 1M input and $0.21/€0.20 per 1M output tokens.
- [SSH access to the Gclaw VM](https://docs.gcore.com/gclaw/ssh-access.md): Obtain SSH access to a Gclaw dedicated VM by generating an SSH key pair, asking the Gclaw assistant to authorize the public key, and connecting as the ubuntu user via the VM IP address.