AI-powered moderation for WhatsApp groups
Proudly powered by NanoClaw
Website · Support Group · Deploy · Docs
git clone git@github.com:TomGranot/groupguard.git
cd groupguard
./setup.shOr use Claude Code for guided setup: run claude then /setup.
Requirements: Node.js 20+, Docker, Claude Code
GroupGuard sits in your WhatsApp groups and enforces rules automatically. Beyond moderation, it's a full Claude assistant — it can answer questions, search the web, schedule tasks, and manage files. The moderation runs silently in the background.
@GroupGuard enable no-spam and no-links for this group
@GroupGuard set observation mode — log violations but don't delete
@GroupGuard add a keyword filter blocking "crypto" and "forex"
@GroupGuard show moderation stats for the last week
@GroupGuard what's the weather in Tel Aviv?
14 built-in moderation rules. Each group gets its own configuration — mix and match as needed.
Control what format of messages are allowed.
| Guard | What it does |
|---|---|
text-only |
Only text messages allowed — blocks images, videos, stickers, voice notes |
media-only |
Only media allowed — blocks plain text messages |
video-only |
Only video messages allowed |
voice-only |
Only voice notes allowed |
no-images |
Blocks images and photos |
no-stickers |
Blocks stickers |
Filter messages based on their content.
| Guard | What it does |
|---|---|
no-links |
Blocks any message containing a URL |
no-forwarded |
Blocks forwarded messages — only original content allowed |
max-text-length |
Blocks messages longer than a limit (default: 2000 characters) |
keyword-filter |
Blocks messages matching specific words or regex patterns — configure your own blocklist |
Rate limiting and access control.
| Guard | What it does |
|---|---|
no-spam |
Blocks rapid-fire messages — more than 5 messages in 10 seconds triggers it |
slow-mode |
Limits users to one message every N minutes (default: 5) |
quiet-hours |
Blocks messages during set hours (default: 10pm–7am) |
approved-senders |
Whitelist mode — only approved users can send messages |
All guards are configurable. Defaults are sensible, but you can tune thresholds, time windows, keyword lists, and approved users per group.
Message arrives → Is sender an admin? → Yes: skip (admins are exempt)
→ No: run guards
→ All pass: message stays
→ Guard blocks: delete + DM sender with reason
Observation mode lets you test rules without deleting — violations get logged but messages stay. DM cooldown prevents notification spam (one DM per user per 60 seconds). Everything is logged to SQLite.
Talk to your bot with the trigger word (default: @GroupGuard):
@GroupGuard enable no-spam for this group
@GroupGuard show me the last 10 moderation violations
@GroupGuard schedule a daily summary at 9am
@GroupGuard remind me every Monday to check group stats
From the admin channel, you control everything:
@GroupGuard list all groups and their guard configs
@GroupGuard enable observation mode for Work Team
@GroupGuard register the "Family Chat" group
Tell Claude Code what you want:
- "Add a guard that blocks messages with more than 3 emojis"
- "Change the DM message when a message is blocked"
- "Add a daily moderation report to the admin group"
Or run /customize for guided changes. The codebase is small enough that Claude can safely modify it.
Works locally or on a server. See the deployment guide for Hetzner, DigitalOcean, and other options starting at $4/month.
Single Node.js process. Guards run on the host for instant enforcement. Agent responses run in isolated Docker containers. Per-group message queues. Full details in docs/SPEC.md.
MIT