Features & Capabilities

Integrations at a glance

OpenClaw connects to 50+ channels, skills, and model providers. Browse by category:

• Channels — WhatsApp, Telegram, Discord, Slack, Signal, iMessage, Teams, WebChat, Matrix, Zalo, and more
• Skills — ClawHub registry (5,000+ skills), productivity, dev, media, channels; install with npx clawdhub@latest install <skill>
• Model providers — Anthropic, OpenAI, Ollama, Gemini, and 40+ others; tiered routing and failover

• Gateway WebSocket management layer providing session management, presence tracking, configuration control, scheduled tasks, webhook support, control interface, Canvas hosting, and a Web UI with token usage dashboard
• Command-line interface: gateway management, agent control, message sending, Setup Wizard, and diagnostic commands
• Agent runtime environment operating in RPC mode with support for tool streaming and block-based streaming
• Session architecture: main session for direct conversations, isolated group sessions, configurable activation modes, queue management, and reply handling
• Proactive and always-on: The agent can run on a schedule (heartbeats, cron) so it “wakes up” to run tasks, briefings, or checks without you starting a chat—see Automation. You can send a request and keep working while the agent runs in the background (non-blocking).
• Media processing system: image, audio, and video handling with transcription integration, file size limits, and temporary file management

Browser Automation

Dedicated Chrome/Chromium browser instance with comprehensive automation features:

• Automated web navigation and page interaction
• Form completion and automated submission
• Website data extraction and parsing
• Screen capture and snapshot generation
• Automated file upload functionality
• Browser profile configuration and management

Interactive Canvas

Dynamic visual workspace controlled by the agent using A2UI framework:

• A2UI interface manipulation with push and reset capabilities
• In-browser JavaScript code execution
• Canvas state capture and snapshot functionality
• Visual interaction environment for agent operations

Nodes

Device-level capabilities:

• Camera snap/clip
• Screen recording
• location.get for location services
• System notifications
• macOS-only: system.run and system.notify

Cron & Automation

• Cron jobs: Schedule recurring tasks
• Wakeups: Scheduled wake-up calls for the agent
• Webhooks: External triggers via HTTP endpoints
• Gmail Pub/Sub: Real-time email triggers

Skills Platform

• Bundled, managed, and workspace skills
• Install gating + UI
• Community skill registry (ClawHub)
• Self-improving: can write its own skills

Voice Wake

Always-on speech recognition for macOS/iOS/Android:

• Continuous listening
• Wake word detection
• Push-to-talk support
• Multi-language support

Talk Mode

Continuous conversation with ElevenLabs text-to-speech:

• Voice responses
• Audio message support
• Custom voice profiles
• Natural conversation flow

Persistent Memory

• Remembers you and becomes uniquely yours
• Your preferences, your context, your AI
• Context persists 24/7 across sessions

Memory Files

• Daily notes formatted in Markdown
• Auto-generated each day
• Plain text log of interactions
• Searchable and editable
• Integrate with Obsidian, Raycast, or Hazel

Skills Platform

• Extend with community skills or build your own
• Self-improving: can write its own skills
• Ask it to add functionality, and it creates the code
• Skills stored in workspace directory

AI Agent Community

Connect your OpenClaw agent to Moltbook, a social network for AI agents where they can share experiences, discuss topics, and form communities with other AI agents. Your agent can also monetize its capabilities through ClawTasks, a bounty marketplace where agents post and complete tasks for USDC.

ClawHub

The skill dock for sharp agents—the community repository for OpenClaw skills with 5,000+ community-built skills available.

• Browse and search 5,000+ skills organized by category
• Install skills with one command: npx clawdhub@latest install <skill-name>
• Agent can search for skills automatically
• Version skills like npm packages
• Community-contributed skills from developers worldwide
• Resources: ClawHub.ai, Awesome OpenClaw Skills

Full System Access

• Read and write files
• Run shell commands
• Execute scripts
• Set up cron jobs
• Automate workflows
• Full access or sandboxed—your choice
• Can improve itself by modifying its own configuration

Browser Control

• Browse the web
• Fill forms
• Extract data from any site
• Full automation capabilities
• Dedicated Chrome/Chromium instance

macOS App

• Menu bar control plane
• Voice Wake/PTT
• Talk Mode overlay
• WebChat + debug tools
• Remote gateway control over SSH

iOS Node

• Pairs as a node via the Bridge
• Canvas surface
• Voice Wake
• Talk Mode
• Camera and screen recording
• Bonjour pairing

Android Node

• Pairs via the same Bridge + pairing flow as iOS
• Exposes Canvas, Camera, and Screen capture commands
• Talk Mode support
• Optional SMS support

macOS Node Mode

• system.run/notify exposure
• Canvas/camera access
• System-level integrations

• Multi-agent routing: Route inbound channels/accounts/peers to isolated agents
• Workspaces: Per-agent sessions and isolation
• Session tools: sessions_list, sessions_history, sessions_send for agent-to-agent communication
• Group routing: Mention gating, reply tags, per-channel chunking