Where links live
A domain you own, or the shared domain on the Free plan. Add it, publish a CNAME and a TXT record, and verify it.
Platform
A route is a domain and path you own, like go.acme.dev/launch. Point it at as many targets as you want and choose how traffic gets split. The link in the wild never changes. Only where it sends people does.
The model
A domain you own, or the shared domain on the Free plan. Add it, publish a CNAME and a TXT record, and verify it.
A slug on a domain, plus a selection method. Create as many as you like. Routes are unlimited on every plan.
A destination URL with a weight or a match value. Targets are unlimited too, and you can disable one without deleting it.
Selection methods
Every route uses one. Pick a method to see exactly what we check and how a visitor ends up at a target.
If a visitor matches no target, the route sends them to its default target, picks one at random, or returns an error. You choose which.
API
Create domains, routes, and targets. Change weights. Pull risk scores and audit events. Issue API keys for your org. It's plain HTTP, and the API is on every plan, Free included, with a rate limit of 90 requests a minute on Free and 180 on paid plans.
Official SDKs for Go, Python, and JavaScript are coming soon.
# Ramp a canary from 10% to 50% API=https://api.divert.ly/v1 ROUTE=$API/domains/$DOMAIN_ID/routes/$ROUTE_ID curl -X PUT $ROUTE/targets/$CANARY_ID \ -H "Authorization: Key $DIVERTLY_KEY" \ -H "Content-Type: application/json" \ -d '{ "weight": 50 }'
Your domain
Add your link domain, publish the two DNS records we give you (a CNAME and a TXT record), and verify. Your DNS stays with your current provider and your links live on your brand. Full nameserver delegation is coming soon.
Free uses the shared dvt.fyi domain. Dev includes one custom domain, Team five, and Scale ten.

Security
A link you can repoint by API is exactly what an attacker wants. So every domain and every target is checked before it carries traffic, and checked again after.
A domain can't serve a single redirect until you've published a DNS record we issue and we've verified it. We keep re-verifying after that.
Each one is scored on several independent signals, including threat-intelligence blocklists, when you add it, when you change it, and on a recurring schedule. You can read the verdict, the score, and the reason through the API.
Scoring on the way in isn't enough, because plenty of links are clean on day one and turn on day thirty. A flagged target is pulled out of rotation at the redirect itself, and the rest of the route keeps working.
New accounts are checked against known-bad signals before they're created, so abuse has a harder time getting started on shared infrastructure.
Our own domains, and sensitive categories like government, military, and education domains, can't be added as custom domains. That closes off the most convincing impersonation targets.
Activity in your organization is recorded as audit events you can read through the API, so when something looks wrong you can see how it got that way.
Phishing, malware, spam, impersonation, or anything else that breaks the rules. Tell us which link and what's wrong. We review every report and act on the links that violate our terms.
Also included
The routing is the headline. These come with it, on every plan.
Click metrics for every route, with 30 days of retention.
Generate a QR code for any route through the API. The code points at the route, so repointing the link doesn't mean reprinting.
Issue and revoke keys for your organization, and read the audit events behind them.
Create a route, add two targets, and watch the traffic split. It takes about five minutes.