Powerful Features
Everything you need to interact with AI agents productively.
Channel → Session → Thread
Organize your agent interactions through a three-layer hierarchy that keeps complex workflows clean and navigable.
- Channels — one workspace per agent, listed in the sidebar
- Sessions — tabbed conversations for parallel work
- Threads — branch from any message without cluttering the main chat
From text to clickable UI — automatically
Agent responses are rendered as interactive elements — buttons, radio groups, and selection cards. Multi-step workflows feel like a guided wizard, not a wall of text you have to parse and retype.
- Clickable option buttons and radio selectors
- Inline model switcher for on-the-fly changes
- Skill command bar for quick actions (
/model,/help,/skills)
Schedule agents to work while you sleep
Set up cron-style background tasks that run your agents on autopilot. Each task has its own prompt, schedule, model, and complete execution history.
- Configurable cron schedules with one-click enable/disable
- Full execution history with job-level detail
- Run any task on demand with "Run Now"
Your conversations, only you can read
Enable E2E encryption with a single password — set the same password in the web UI and in your OpenClaw plugin. All sensitive data is encrypted locally with AES-256-CTR before it ever touches the network.
- Zero-knowledge server — stores only ciphertext, no keys or salts
- Zero overhead — ciphertext is the same size as plaintext
- Chat messages, cron prompts & job summaries — all encrypted
- Database stores BLOB — even with full DB access, content is unreadable
Real-time debug log, built right in
A collapsible debug panel gives you live visibility into WebSocket events, task loading, agent responses, and errors — all without leaving the chat.
- Filter by level: ALL, WS, API, INF, WRN, ERR
- Real-time streaming of gateway events
- Diagnose connection & agent issues instantly