Skip to navigation

Risk types and codes

What the four-digit codes in the Fraud API’s riskTypes field mean, and how to block on them.

Fraud API v3 returns a riskTypes field next to probability. Each entry is a four-digit integer code for a type of invalid traffic (IVT) in Pixalate’s IVT Taxonomy 2.0. The codes tell you why an IP address, device ID or user agent is flagged, and you can use them in blocking rules.

{
"probability": 0.92,
"riskTypes": [2502, 2201]
}

For the request itself, see the Fraud API guide and the API reference.

How riskTypes behaves

  • It lists the risk types associated with all the values you sent (ip, deviceId, userAgent) during the look-back period, so it can contain more than one code.
  • It’s null when probability is 0 (not flagged) or -1 (whitelisted).
  • It’s null when a value is flagged but no risk type is available for it.
  • It’s never an empty array, and each code appears at most once.

Code format

Each code follows the format T C NN:

PartMeaningValues
TIVT class1 = GIVT (general invalid traffic), 2 = SIVT (sophisticated invalid traffic)
CCategory within the classSee the tables below
NNType within the category01, 02, …

For example, 2201 is Invalid Proxy Traffic: class 2 (SIVT), category 2 (False Representation), type 01.

Block by class, category or type

Because the code is structured, you can match at any level:

RuleMatchesCheck
All SIVT2xxxcode // 1000 == 2
One category, such as Known Dangerous Sources25xxcode // 100 == 25
One type, such as Known Dangerous or Fraudulent IPv4 Addresses2502code == 2502

You can use these rules on their own or together with a probability threshold. See Decide when to block.

Categories

Code prefixClassCategory
11xxGIVTIrregular Pattern
12xxGIVTKnown Crawlers
13xxGIVTData Center
14xxGIVTInvalid Placements
21xxSIVTAutomated Browsing
22xxSIVTFalse Representation
23xxSIVTIncentivized Human Invalid Activity
24xxSIVTMisleading User Interface
25xxSIVTKnown Dangerous Sources
26xxSIVTSupply Chain

Codes

The tables list every type in IVT Taxonomy 2.0. Types marked Upcoming are defined in the taxonomy but aren’t live yet. A lookup returns only the codes associated with the values you sent.

GIVT: Irregular Pattern (11xx)

CodeNameDefinition
1101Repeat TransactionsImpressions with the same ID and attributes generated multiple times; usually an integration issue.
1102Repeat Click TransactionsClicks with the same ID and attributes generated multiple times; usually an integration issue.
1103Spoofed User AgentsTraffic from a browser identifier that presents more than two user agents within 24 hours.
1104Fraudulent User AgentsTraffic from user agents that match no known browser or app.
1105Missing User AgentsImpressions whose request carries no user agent.
1106Auto Refresh FraudImpressions from a single source arriving at intervals too regular to be human.
1107Reserved Network IPsImpressions whose IP address is in private or reserved network space.
1108Fraudulent ClicksClicks recorded too soon after their impression for a person to have reacted.
1109Invalid Prefetch TrafficImpressions from requests self-declared as prefetch or preview.
1110Non-Rendering Device TrafficTraffic from devices that can’t render ad content.

GIVT: Known Crawlers (12xx)

CodeNameDefinition
1201IAB CrawlersTraffic from user agents on the IAB/ABC International Spiders & Bots List.
1202Declared AI Agents (Upcoming)Traffic from AI agents and LLM crawlers that self-identify in the user agent.

GIVT: Data Center (13xx)

CodeNameDefinition
1301Data Center TrafficTraffic from known data centers and hosting providers that generate invalid traffic, excluding legitimate human activity routed through VPNs or virtual machines.

GIVT: Invalid Placements (14xx)

CodeNameDefinition
1401Invalid Placement TrafficImpressions served to illogical placement sizes, such as 0x0 or 1x1.

SIVT: Automated Browsing (21xx)

CodeNameDefinition
2101Infected Device TrafficImpressions generated by a single device at a statistically inflated rate.
2102Infected Device SessionsTraffic from devices with a statistically inflated number of sessions per hour.
2103Headless Browsers or Emulators (Upcoming)Traffic from headless browsers or emulated devices posing as physical hardware.
2104Known Automation Systems (Upcoming)Traffic from CTV devices whose behavior matches a device farm.

SIVT: False Representation (22xx)

CodeNameDefinition
2201Invalid Proxy TrafficTraffic from intermediary proxies that inflate traffic counts, pass invalid traffic, or break protocol.
2202Data Center Proxy TrafficTraffic from proxies on data center or hosting IPs that inflate counts, pass invalid traffic, or break protocol, excluding validated SSAI.
2203IP Address MismatchImpressions whose declared IP and ISP differ from the IP and ISP Pixalate detects.
2204Spoofed iCloud Private RelayImpressions whose declared IP claims iCloud Private Relay but differs from the IP and ISP Pixalate detects.
2205Fraudulent XFFImpressions whose X-Forwarded-For header reports a user IP that differs from the IP the ad renders to.
2206Fraudulent Device IDsTraffic from device IDs seen across a statistically inflated number of IPs, apps, platforms, or locations.
2207App MisrepresentationImpressions whose declared bundle ID differs from the app detected by Pixalate.
2208OS MisrepresentationImpressions whose declared operating system differs from the OS detected by Pixalate.
2209Domain MisrepresentationImpressions whose declared domain differs from the domain detected by Pixalate.
2210User Agent Bundle ID Device MismatchImpressions whose declared operating system is inconsistent with the bundle ID format, user agent, or device.
2211Falsified GeolocationImpressions whose declared GPS location deviates significantly from the IP reverse location.
2212Cookie StuffingTraffic from a single browser identifier recycled across a statistically inflated number of IP addresses.
2213Spoofed ReferrerTraffic from a single source reaching one domain through a statistically inflated number of referrer domains.
2214Reserved (Upcoming)Not yet live.
2215SDK Spoofing (Upcoming)Impressions whose ad request claims an SDK that didn’t initiate the transaction.
2216Reserved (Upcoming)Not yet live.

SIVT: Incentivized Human Invalid Activity (23xx)

CodeNameDefinition
2301Click FarmsTraffic from IPs where a statistically inflated number of devices generate coordinated clicks.
2302Infected Device ClicksClicks generated by one device at a statistically inflated rate.
2303Falsified Clicks (Upcoming)Clicks recorded after an app download has begun but before the install completes.

SIVT: Misleading User Interface (24xx)

CodeNameDefinition
2401Obfuscated Ad ServingTraffic from domains or apps Pixalate identified for stacked, hidden, or invisible ad placements.

SIVT: Known Dangerous Sources (25xx)

CodeNameDefinition
2501Known Dangerous or Fraudulent DevicesTraffic from device IDs Pixalate identified as dangerous or fraudulent.
2502Known Dangerous or Fraudulent IPv4 AddressesTraffic from IPv4 addresses Pixalate identified as dangerous or fraudulent.
2503Known Dangerous or Fraudulent IPv6 Addresses (Upcoming)Traffic from IPv6 addresses Pixalate identified as dangerous or fraudulent.
2504Known Dangerous or Fraudulent DomainsTraffic from domains Pixalate identified as dangerous or fraudulent.
2505Defased AppsTraffic from apps delisted from their respective app stores but still serving ads.
2506Adware, Malware or PhishingTraffic from domains associated with adware, malware, or phishing.

SIVT: Supply Chain (26xx)

CodeNameDefinition
2601SCO Publisher is NullImpressions whose first SupplyChain node doesn’t identify as a publisher.
2602SCO Unauthorized SellersImpressions whose first SupplyChain node isn’t listed as DIRECT in ads.txt or app-ads.txt.

Handle new codes

Pixalate may add types to the taxonomy. Write your rules so that an unfamiliar code doesn’t break your integration: match on the class or category prefix where you can, and log codes you don’t recognize.