A VPN means something different on a newsletter signup than on a withdrawal. Set what every signal is worth, per platform, and where high risk begins. The change applies to the next request scored.
A scoring policy is a set of weights and a threshold, kept separately for each platform.
What each signal adds to the score.
Any whole number from 0 to 10,000. Signals you do not set keep the shipped default. A zero turns the contribution off while the trigger stays on the record.
Where high risk begins.
The score at which a request is high. Medium begins at half of it. The default is 30, roughly two strong network signals or one plus a device signal.
Price by method, or by confidence.
VPN and residential proxy detections can be weighted per detection method, or as one signal priced by how sure the detection is. One mode applies at a time, so a finding is never counted twice.
A stricter rule where it matters.
The organisation’s policy applies to every key. A key can carry its own policy through the API, so checkout can run a stricter rule than the marketing site without a second account.
Fingerly never blocks a request. An example of how an application might map levels to actions at two different moments.
| Level | Score at the default threshold | At signup | At a withdrawal |
|---|---|---|---|
| Low | Below 15 | Create the account | Process it |
| Medium | 15 to 29 | Verify the email first | Ask for a second factor |
| High | 30 or more | Hold for review | Refuse and notify |
An example application’s policy. Start in monitoring mode and measure how legitimate customers are affected before enforcing.
Weights are production configuration, and they are handled like it.
Every member can read the policy. Only an owner or an admin can change it.
A save carries the revision it was based on. If someone else saved in between, the change is refused instead of overwriting theirs.
Every assessment keeps the threshold and the policy revision it was scored under, so past decisions do not change meaning.
Return a platform to the shipped values in one step, or remove a key’s override so it follows the organisation again.
Yes. A key can carry its own weights and threshold for each platform through the API, overriding the organisation’s policy for that key alone. Removing the override makes the key follow the organisation again.
On the next request scored. There is no cache to wait for.
No. A weight for a signal a platform cannot produce, such as a rooted device on the web, is refused rather than silently accepted and never used.
Start with the signals, keep your own decisions, and pay only for what you identify. No credit card needed.