SidekickJoin the waitlist
While you slept, your agent shipped the work.
Sidekick is a cloud AI agent that reads your email, monitors your repos, triages your Slack, and acts — on your schedule, at your autonomy level. Connect your tools once. Never babysit it again.
[ SETUP SEQUENCE / 4 STEPS ]
Pick from Gmail, Slack, GitHub, Notion, Linear and more. One OAuth flow per service — no API keys to paste, no config files.
Choose ask-first, act-then-report, or fully-autonomous — per integration. Define quiet hours, priority rules, and which actions always need your sign-off.
Search by natural language — 'help me schedule meetings' surfaces Calendar & Scheduling skills. One-click install maps env vars to your OAuth connections automatically.
The agent loop gathers events, loads your context, reasons with Claude, and acts — on a schedule you control. Decisions are logged and auditable. You sleep. It ships.
[ APPROVALS / ASK-FIRST MODE ]
The approval queue puts you in the loop before anything consequential happens. Sidekick shows you exactly what it wants to do and why — then waits for your word.
Your agent is ready. Are you?
Join the early access list. We'll reach out when your slot opens. Questions first? Email us at sf-core-org-support-sidekick@saas-factory.ai
[ DASHBOARD / LIVE ]
The dashboard shows every decision your agent makes, every action it queued, and every integration it touched — real-time, with full reasoning chains. You're always in control, even when you're not watching.
[ INTEGRATIONS / ACTIVE ]
Sidekick normalises data from every connected tool into a unified event stream — so your agent can reason across Gmail, Slack, GitHub, Linear, Notion, Outlook, Telegram, WhatsApp, Discord, and Google Calendar as one coherent picture.
Reads threads, drafts replies, surfaces messages that need a human decision. Push notifications mean zero polling lag.
[ ACTIVITY / REAL-TIME ]
The agent decision transparency log shows exactly what triggered each action, what the agent reasoned, and what it did. Not a black box — a complete audit trail for every move your AI makes.
Real-time activity feed See trigger → reasoning → action for every agent decision as it happens
Approval queue Ask-first mode surfaces pending actions before they execute — approve, edit, or reject with one tap
Destructive action gate Deleting data, emailing new contacts, or revoking access always requires your explicit confirm
[ MEMORY / THREE-TIER ]
Working memory keeps recent context sharp. Episodic memory stores daily summaries. Semantic memory uses vector search for long-term pattern recall. You never re-explain yourself.
Recent events held in Redis for instant in-context recall during the agent loop.
Daily summaries stored in Postgres — the agent knows what happened yesterday, last week, last quarter.
[ ANALYTICS / USAGE ]
[ SKILLS / CLAWHUB ECOSYSTEM ]
Browse the ClawHub marketplace directly from your dashboard. Search with natural language — embeddings match intent to capability, not keywords. Install in one click; env vars map to your OAuth connections automatically.
Type 'help me prep for meetings' and find the right skill — even without exact keyword matches.
[ CONTROL / TRUST LAYER ]
Ask-first, act-then-report, or fully autonomous — configurable per integration with read-only mode for sensitive accounts.
Define no-action windows with timezone awareness. Events queue silently and flush when quiet hours end.
[ FAQ / COMMON QUESTIONS ]
Real-time activity feed with trigger → reasoning → action trail
One-click Emergency Kill Switch halts everything instantly
Connection health, last sync time, and permission scopes visible at a glance
Monitors channels, responds to mentions, and routes important threads to your approval queue before replying.
Triages issues, reviews PR notifications, tracks CI status, and posts sprint summaries without leaving your chat.
pgvector embeddings for long-range pattern recall — 'how does Sarah usually respond to rescheduling requests?' answered from history.
The analytics dashboard shows actions taken per day, response times, most active integrations, memory utilisation, and skill usage breakdown. Know exactly where your agent is spending cycles — and where to point it next.
Daily action counters vs. plan limit shown in the dashboard header
Per-integration rate limits prevent runaway API costs
Skill usage breakdown — see which automations pull their weight
Write natural-language instructions, pick your integrations, test in sandbox, export as SKILL.md for sharing.
Already using OpenClaw? Export your skills, import to Sidekick, connect your accounts — migration guide included.
All skill credentials stored with AES-256-GCM encryption at rest, with automatic key rotation support.
One button halts all agent activity, cancels queued actions, and disconnects integrations. Instantly, unconditionally.
