Help Center← Back to Dashboard
Getting Started
What is Aithroyz?Quickstart: First EnvironmentCloud CredentialsPlans & Approvals
Environments
OverviewLifecycle PhasesTTL Auto-DestroyExtending TTLDestroying an Environment
Tools Reference
OverviewElastic Stack (SIEM)Wazuh (XDR)MITRE CalderaTheHive & DFIR-IRISVelociraptorOpenCTIGrafana + PrometheusShuffle SOARn8nUptime KumaLLM GatewayOpen WebUIFlowiseOpenClawOllamaQdrantLangfusePortainerGiteaSonarQubeCode ServerMattermostMinIOMetabaseHashiCorp VaultKeycloak SSONetBoxLocalStack
Access & Security
Google SSOTenant IsolationPasskeys & MFATeam Members
API & Integrations
API KeysMCP Tools (Clevername)Terraform ExportWebhooks & Callbacks
Stack Presets
SOC PlatformIR / DFIR LabThreat HuntingQuick Sandbox
Settings
Cloud KeysAPI KeysBillingAudit Log
Troubleshooting
Common IssuesDeployment FailuresDNS & ConnectivityTool Health Checks
Aithroyz Help
Help CenterTools ReferenceOpenClaw

OpenClaw (Multi-channel AI Gateway)

OpenClaw is a self-hosted AI gateway that connects messaging apps — Telegram, Discord, WhatsApp, Slack, Signal — to a private AI agent (Pi) with tool use, memory, and automation.

Aithroyz pre-seeds the gateway token and wires the LLM Gateway (or Clevername Hub) as Pi's AI provider at provisioning time — OpenClaw is ready to connect a channel immediately after login.

Access

URL: https://openclaw.<env-name>.ops.aithroyz.com
Auth: Paste the gateway token from the Credentials panel on first load. The token is the only credential — keep it private.

What OpenClaw does

Pi agent
A persistent AI agent that responds to messages in any connected channel. Pi has memory across conversations, can use tools, and run scheduled tasks autonomously.
Skills (5400+)
Installable capabilities from the ClawHub marketplace — web search, email, calendar, code execution, API calls, and hundreds of integrations.
Channels
Telegram, Discord, WhatsApp, Slack, Signal, iMessage, Matrix, and more. Each channel is added independently; Pi responds in all of them simultaneously.
Cron scheduler
Built-in task scheduler for autonomous actions — have Pi send a daily briefing, scrape a site, or run a report on a fixed schedule.

Connecting a channel

Go to the Config tab and add a channel block in JSON. Here is an example for Telegram:

{
  "channel": "telegram",
  "token": "<your-telegram-bot-token>",
  "allowFrom": ["123456789", "987654321"]
}

The allowFrom array restricts which Telegram user IDs can talk to Pi. Leave it empty to allow all users, but this is not recommended for private deployments.

✓
Get your Telegram bot token by messaging @BotFather on Telegram. Your user ID can be found by messaging @userinfobot.

After saving the config, OpenClaw registers a webhook with Telegram automatically — no polling or external tunnels required.

Installing skills

Skills extend Pi's capabilities with new tools it can call during conversations:

1
Open the Skills tab
Click Skills in the left sidebar to open the ClawHub skill browser.
2
Search or browse
Filter by category (Productivity, Developer, Research, etc.) or search by name.
3
Install
Click Install on any skill. Some skills require an API key — enter it when prompted and it is stored encrypted on your instance.
4
Test in chat
Ask Pi to use the skill by name in any connected channel. Pi will invoke it automatically when relevant.

Tips

allowFrom whitelist
Always set allowFrom in each channel block in production. Without it, any user who finds your bot can interact with Pi and consume your LLM quota.
Workspace persistence
Pi's memory, notes, and uploaded files are stored at /opt/openclaw/workspace on the VM. This directory persists across container restarts.
LLM Gateway pairing
When the LLM Gateway is in the same plan, Pi routes all completions through it — giving you centralized token logging, rate limits, and model switching.
Documentation
Full channel config schemas and skill authoring guide at docs.openclaw.ai.
Related Articles
LLM GatewayRead article →Flowise (LangChain Visual Builder)Read article →n8n (Workflow Automation)Read article →