OpenClaws Documentation

What is OpenClaws?

OpenClaws is a managed hosting platform for OpenClaw, the open-source AI assistant with 260,000+ GitHub stars. Each subscriber gets their own dedicated AI gateway — a 24/7 assistant running on isolated infrastructure with persistent memory, 1000+ tool integrations, and 8 messaging channels.

Your gateway runs on its own Fly.io VM with dedicated CPU, RAM, and persistent storage. No data is shared between users. For deep dives into OpenClaw itself, see docs.openclaw.ai.

Quick Start

1. 1

   Sign in

   Go to openclaws.biz/login and sign in with Google or X (Twitter).

2. 2

   Subscribe ($29/month)

   Navigate to Settings → Billing and click "Subscribe." Your personal AI gateway deploys automatically (~30 seconds).

3. 3

   Open your Control UI

   Click "Open your OpenClaw" on the Dashboard. This opens your gateway's Control UI at your-name.openclaws.biz. You'll be auto-authenticated.

4. 4

   Start chatting

   Use the Chat tab in the Control UI to talk directly to your AI assistant, or connect a messaging channel (Telegram, Discord, WhatsApp, etc.).

Control UI Overview

The Control UI is the browser-based dashboard for managing your gateway. It's organized into 4 groups with 13 tabs:

Accessing the Control UI

Your Control UI is at https://your-slug.openclaws.biz. Click "Open your OpenClaw" from the Dashboard to open it with auto-authentication. The gateway token is passed automatically via the URL hash.

Device Pairing (First Connection)

New browser connections require a one-time pairing approval. If you see "pairing required," the gateway needs to approve your device. When you use the "Open your OpenClaw" button, the gateway token is included automatically — but you may still need to approve the device on first connection. Contact support if you're stuck on pairing.

Each browser profile gets a unique device ID. Clearing browser data will require re-pairing. See OpenClaw docs: Device Pairing for details.

Chat

The Chat tab provides a direct conversation interface with your AI assistant. Key features:

• •Message thread with streaming responses and elapsed timer
• •Session selector — switch between conversation contexts or start new ones
• •Thinking toggle — show/hide the AI's reasoning process
• •Focus mode — fullscreen chat with all UI chrome hidden
• •Image paste support — paste images directly from clipboard
• •Message queue when assistant is busy processing
• •Stop button to abort long-running responses

Messaging Channels

Connect your AI assistant to 8 messaging platforms so you can talk to it from anywhere. Each channel has its own configuration section in the Channels tab.

Each channel shows live status (Configured, Running, Connected) and health metrics. Use the Probe button to test connectivity.

Cron Jobs

Schedule recurring agent runs from the Cron Jobs tab. Great for morning briefings, automated reports, monitoring, or any task you want to run on a schedule.

Creating a Job

• 1.Name & Agent — Give the job a name and optionally pick a specific agent
• 2.Schedule — Choose Every (e.g., every 30 minutes), At (specific datetime), or Cron (cron expression with timezone)
• 3.Execution — Run in the Main session (posts a system event) or Isolated (dedicated agent turn with your prompt)
• 4.Delivery — Announce posts a summary to chat, Webhook POSTs results to a URL, or None keeps it internal

The run history panel shows all past executions with status, duration, and output. You can enable/disable, manually run, or delete individual jobs.

Skills

Your gateway comes with 51 built-in skills — pre-packaged capabilities like GitHub, Notion, Slack, Discord, image generation, PDF editing, text-to-speech, and more. Manage them from the Skills tab.

Skill Status

• ●Eligible — ready to use, all dependencies met
• ●Blocked — missing requirements (binary, API key, config, or OS)

Enabling Skills

For blocked skills, the UI shows exactly what's missing. Common fixes:

• •Missing API key — paste your key in the skill's "API key" field and click Save
• •Missing binary — click the "Install" button (e.g., "Install GitHub CLI (brew)")
• •Missing config — configure the related channel or plugin first

Popular Skills

Agents

The Agents tab manages agent workspaces — isolated contexts with their own identity, tools, skills, and model configuration.

• •Overview — workspace name, identity (name, emoji), default status
• •Model Selection — set primary model and comma-separated fallbacks
• •Files — list and edit agent workspace files
• •Tools — tool policy editor and profiles
• •Skills — per-agent skill allowlist
• •Channels — per-agent channel configuration
• •Cron Jobs — per-agent scheduled tasks

For multi-agent architectures, see OpenClaw docs: Multi-Agent.

Sessions

The Sessions tab shows all active conversation sessions with per-session overrides:

• •Thinking — off / minimal / low / medium / high (controls how much reasoning the AI shows)
• •Verbose — inherit / off / on / full (controls response detail level)
• •Reasoning — inherit / off / on / stream (controls reasoning output mode)

Each session shows its key, label, kind (direct/channel), last update time, and token usage against the context limit (e.g., 14,809 / 200,000). You can delete individual sessions to free resources.

Usage Monitoring

The Usage tab provides comprehensive token and cost analytics:

• •Date range filters — Today, 7d, 30d, or custom range
• •Token vs Cost view — toggle between raw token counts and estimated costs
• •Activity timeline — visualize when your gateway is busiest
• •Session breakdown — per-session token usage, error counts, message counts
• •Advanced filters — filter by session key, model, provider, agent, channel
• •Export — export usage data for external analysis

Nodes & Devices

The Nodes tab manages execution security and device access.

Exec Approvals

Control what commands the AI can execute on its own:

• •Deny — block all exec commands (most secure)
• •Allowlist — only pre-approved commands can run
• •Full — allow all commands (least secure)

The "Ask" policy controls whether the UI prompts you for approval: Off (never ask), On miss (ask for unknown commands), Always.

Device Management

View paired devices, rotate or revoke device tokens, and manage pending pairing requests. Each device has scoped roles (operator, admin, read, write, approvals, pairing).

Configuration

The Config tab provides a full settings editor for your gateway. It edits ~/openclaw/openclaw.json with 25+ categories:

Use the search bar to find specific settings. Toggle between Form view (structured UI) and Raw view (JSON editor). Changes are validated before saving. See OpenClaw docs: Configuration for all available options.

Debug & Logs

Two tabs for troubleshooting:

• •Debug — Raw JSON snapshots (status, health, heartbeat), manual RPC calls to the gateway, model catalog
• •Logs — Live JSONL log tail with 6 log levels (trace/debug/info/warn/error/fatal), text search, auto-follow, and export

Connecting Toolkits

OpenClaws integrates with 1000+ services through Composio. Toolkits are OAuth-connected external services (Gmail, GitHub, Slack, etc.) managed from the OpenClaws dashboard — distinct from the Skills in the Control UI, which are built-in gateway capabilities. When you connect a toolkit, your gateway automatically gets access to it.

1. 1.Go to the Toolkits page in the OpenClaws dashboard
2. 2.Search for the service you want (Gmail, Slack, GitHub, etc.)
3. 3.Click Connect and authorize via OAuth
4. 4.Your assistant can now use that service — try asking it to read your Gmail or send a Slack message

The bridge works automatically: your Composio API credentials and entity ID are injected into your gateway at provisioning time. The gateway's Composio skill discovers connected services and executes actions on your behalf using OAuth tokens managed by Composio.

Choosing a Model

OpenClaws supports multiple AI models. Switch models in Settings → Model Selection on the OpenClaws dashboard, or in Agents → Model Selection in the Control UI:

Bring Your Own API Keys

Use your own API keys for full control over costs and model access:

• •Anthropic API Key — starts with sk-ant-api...
• •Anthropic OAuth Token — starts with sk-ant-oat01-... (from claude setup-token)
• •OpenAI API Key — starts with sk-...

All keys are encrypted at rest using AES-256 via Supabase pgcrypto. Keys are only decrypted server-side when pushed to your isolated gateway.

FAQ

Support

Need help? Join our Discord community or email us at support@openclaws.biz. For OpenClaw documentation, visit docs.openclaw.ai.