Skip to main content
Hashbot Vision protects your Discord server by scanning images, links, and member activity in real time. It combines OCR-based text extraction, AI-powered scam analysis, malicious URL detection, a full audit logging system, and cross-server threat intelligence — all in one bot.

​
What Vision does

OCR scam detection

Extracts text from images posted in your server and scans for scam keywords, phishing attempts, and social engineering tactics — even when the text is hidden inside a screenshot or QR code.

AI analysis

Goes beyond keyword matching with AI-powered contextual analysis that identifies scam patterns, impersonation attempts, and suspicious formatting.

URL scanning

Checks URLs found in messages, images, and QR codes against threat intelligence feeds to flag known malicious links, phishing sites, and wallet drainers.

Action Log

Tracks 27 event types across 8 categories with per-event channel routing, User Insight on member join, and granular ignore rules. Learn more

Vision Insight

Cross-server threat intelligence pooling ban data from servers across the network — 234K+ known threat IDs and growing — alerting or auto-banning known scammers when they join. Learn more

Web dashboard

Configure everything from a visual interface at dashboard.hashbot.com — no slash commands required. Learn more

​
Getting started

1

Invite the bot

Add Hashbot Vision to your server using the invite link. Make sure it has the required permissions (View Channels, Send Messages, Embed Links, Attach Files, Read Message History, Manage Messages, View Audit Log, Kick Members, Ban Members, Timeout Members). View Audit Log is what lets the Action Log show who deleted a message.
2

Set up the Action Log

Run /actionlog-setup to launch the interactive wizard. It walks you through all 8 event categories in under a minute and creates log channels automatically if needed.
3

Enable Vision Insight

Run /vision-insight enabled:True mode:Protect log_channel:#your-channel to start receiving cross-server threat intelligence.
4

Configure OCR detection

Set a log channel with /set-log #channel, enable AI detection with /set-ai enabled, and apply a keyword template with /template apply. Set your punishment with /set-punishment ban.
5

Enable link checking

Turn on automatic URL scanning with /link-checking enabled to catch malicious links in messages and images.
Prefer a visual interface? Skip the slash commands and configure everything from the Vision Dashboard.

​
What happens on a detection

Vision applies your configured punishment (alert, delete, timeout, or ban) and also sweeps the offender’s recent messages, so a multi-message spam burst is cleaned up in one action. If Vision can’t act — for example due to role hierarchy or a missing permission — it posts a “Manual Action Needed” alert naming exactly what’s missing, so your team can follow up.

​
Plans and usage

Vision is free to start for every server. Image scanning is metered per month:
  • Check this month’s usage anytime with /vision usage
  • Upgrade in one step with /vision upgrade (secure Stripe checkout; requires Manage Server)

​
NSFW Protection (Beta)

Vision can also detect explicit images and GIFs posted in chat. It is off by default; turn it on with /nsfw-filter or from the dashboard’s Detection page, and pick what happens on a confirmed match:
  • Alert only (default): posts an alert to your log channel, with the evidence attached behind a spoiler
  • Delete message: removes the message and alerts your mods
  • Delete + timeout: removes the message and times out the poster
Detection is deliberately conservative: images are checked by an image-moderation screen plus an AI judge, only high-confidence explicit matches are actioned, and borderline cases are just logged for review. Age-restricted channels are always skipped. NSFW detection shares your normal scan allowance: the Trial plan includes it within its 500 monthly scans, and paid plans add it with the NSFW variant shown above. One scan covers both OCR and NSFW checks on the same image; animated images use up to 3 scans when NSFW Protection is on. Looking to filter explicit profile pictures instead? That’s NSFW PFP Protection, part of the main Hashbot bot. When a server reaches its monthly cap, image scanning pauses until the 1st of the month — URL checking, QR decoding, and plain-text detection keep running. Vision posts a cap notice with upgrade options to your log channel and the Action Log’s Billing event, and servers with a billing email on file also receive an email alert with upgrade steps. Visit Hashbot Safety Suite to see all available bots.

​
Learn more

Action Log

27 event types, per-event routing, User Insight, and setup wizard

Vision Insight

Cross-server threat intelligence with confidence scoring

Dashboard

Web-based configuration at dashboard.hashbot.com

Commands

Complete slash command reference