Verification & alt detection for Discord

Keep alts, ban evaders, and raiders out. Without the false positives.

Vidar verifies new members through a private DM checkpoint, then scores them with network intelligence, privacy-respecting device fingerprinting, and cross-server identity matching — every decision explained, never a black box.

4-layer
signal model
0
raw IPs stored
< 2s
to a verdict
Verification · @nova
Review
64risk

Top alt match

81%

shares device + network

Identical device fingerprint+2.1
VPN detected+1.1
Account created 2 days ago+0.9
Same /24 network as 1 member+0.7
Has reputation badges−0.6
No raw IP stored · decision is fully explainable
The flow

Verification in four quiet steps

Members never leave Discord to start. Vidar handles the rest and assigns the role you choose.

  1. 01

    Member joins

    Vidar greets the new member with a direct message and an optional hold role.

  2. 02

    DM checkpoint

    They click a one-time verify link that opens a secure checkpoint page.

  3. 03

    Analyze

    Network intel, fingerprint and cross-server matches produce a confidence score.

  4. 04

    Assign role

    Clean members get your verified role instantly; risky ones are held for review.

Capabilities

A complete defense layer

Everything you need to stop evasion and automation — tunable to your server's tolerance.

Alt detection

Cross-references privacy-respecting fingerprints, IP and network hashes to link accounts to known members — flagging ban evaders with a confidence score.

VPN, proxy & Tor intel

Identifies commercial VPNs, public and residential proxies, Tor exit nodes and datacenter IPs, powered by live IP intelligence.

Explainable scoring

Every verdict lists the exact signals and their weights. No opaque numbers — moderators see precisely why a decision was made.

Raid protection

Detects coordinated join-bursts by velocity, account-age clustering and shared signals — tuned so a handful of organic joins never trips the alarm.

Cross-server checks

Optionally consult a global, anonymized sighting index to catch accounts already flagged across other Vidar-protected servers.

Fully configurable

Thresholds, blocked network types, minimum account age, the verified role and raid sensitivity — all from the web dashboard.

The model

Four categories of evidence, one honest score

Vidar accumulates bounded evidence across network, device, account and link signals, then squashes it into a 0–100 risk score. No single heuristic can force a verdict — confident decisions require corroboration.

Read the detection docs

Network

VPN · proxy · Tor · hosting · ASN · geo · provider risk

Device

Hashed fingerprint · UA family · timezone · platform entropy

Account

Age · avatar · badges · MFA · verified email

Link

Shared device · shared IP · same network · banned-alt match

Privacy by design

Powerful detection, minimal data

Vidar is built to be effective and respectful. It derives signals, not dossiers.

No raw IPs

IP addresses are turned into salted, non-reversible hashes the moment they arrive. The plaintext address is never written to the database.

Short retention

Exact-IP hashes are purged on a short schedule; only coarse /24 network hashes are kept longer for correlation.

Data minimization

No message content, no browsing history — only the anonymized facets needed to make a verification decision.

Put Vidar on watch

Invite the bot, pick your verified role, and let the checkpoint handle the rest. Set up takes a couple of minutes.