OpenClaw status dashboard

Why have a Mac mini when you can have a pi? 🥧

Fresh telemetry from the local OpenClaw gateway, synced every 30 minutes. Sessions, tokens, agents, and skills — all in one pulse.

Raspberry Pi is a registered trademark of the Raspberry Pi Foundation.

Active sessions

—

across every agent

Tokens used (recent)

—

sum of the latest sessions

Agents in motion

—

agents reporting activity

Gateway health

—

reachability + auth

Recently active sessions

Snapshot of the running OpenClaw gateway (latest 10 sessions).

refreshed —

    Agents

    Autonomous workers with their own sessions, memory, and model assignments.

    Active Idle

    Skills 0

    Toolkits an agent can invoke — from email to web scraping to image generation.

    Custom 0

    Built-in 0

    Cron jobs & usage

    Auto-updates from status.json. New cron jobs appear automatically after refresh.

    Cron jobs (tracked)

    —

    unique cron IDs seen in sessions

    Recently active cron jobs

    —

    activity in the last 24h

    Cron token usage

    —

    summed across tracked cron sessions

      Glossary

      Key terms used across this dashboard and the OpenClaw ecosystem.

      Session

      A conversation thread between an agent and a user (or system). Each session tracks its own token usage, model, and context window.

      Agent

      An autonomous AI worker with its own config, memory, and skill set. Agents can be triggered by chat, cron, webhooks, or other agents.

      Skill

      A modular toolkit an agent can invoke — like email sending, web scraping, or image generation. Skills are installed from ClawHub or built locally.

      Gateway

      The OpenClaw process that routes messages between channels (Telegram, Discord, etc.), agents, and tools. Think of it as the central nervous system.

      Token

      The unit of text an LLM processes. Roughly 3–4 characters per token. Every message burns input tokens and generates output tokens. Usage is tracked per session.

      Context Window

      The maximum amount of text (in tokens) a model can consider at once. When a session nears its limit, older content gets compacted or summarized.

      Subagent

      A temporary agent spawned by another agent for a specific task (like researching a topic). It runs in isolation and reports back when done.

      Cron Job

      A scheduled task that fires on a timer — can trigger agent turns, shell commands, or system events. Used for things like status updates and email checks.

      LCM

      Lossless Context Management — an OpenClaw plugin that compresses old conversation history into searchable summaries without losing information.

      Model

      The LLM powering a session (e.g., GPT-5.3, Claude, Gemini). Different models have different capabilities, speeds, and costs.

      ClawHub

      The skill marketplace for OpenClaw — search, install, and publish agent skills. Like npm, but for AI toolkits.

      Heartbeat

      A periodic check-in where the agent proactively looks at email, calendar, or other data sources without being asked.