VPX Global Insight +

Telecom intelligence live on our own network - 14-dimension route quality, AIT detection and per-number risk scoring - moving to network-wide, on-chain consensus with the VPX Protocol (Q4 2026). Sub-100ms acknowledgement, routes to 200+ countries.

REST · WebSocket stream · SMPP v3.4 · SFTP batchAPI Protocol
<100ms ack · <2s enriched availability · <250ms AI-scoredResponse Time
GSMA Call Check · STIR/SHAKEN · Ofcom CLI Auth · ITU-T E.164 · our own CLI historyA-Number Sources
HLR · MNP · range/prefix validation · number-allocation registry · premium-range flagsB-Number Sources
Overview

What it
does

VPX Global Insight+ is the intelligence layer over VoicePro's network. Today it powers our own interconnect routing and fraud decisions: real-time route quality scored across 14 dimensions, multi-dimensional anomaly detection, AIT and bilateral-fraud surfacing, and per-number risk scoring, with CLI, HLR, MNP, DNO and TPS lookups on every call.

The differentiation is structural, not algorithmic. Proprietary fraud databases see one carrier's traffic. The VPX design has independent Fraud Nodes corroborate signals across the whole network, breaking the bilateral-trust model that makes AIT undetectable by single operators (industry fraud cost: $39B annually, CFCA 2023). Under that model every blacklist entry, benchmark and anomaly will carry on-chain attestations from independent nodes, so no single operator - including VoicePro Plus - can flag a number unilaterally.

As the VPX Protocol comes online (Q4 2026), the same signals move on-chain and open to enterprise customers via API: a single REST, WebSocket, SMPP or SFTP call will return the full intelligence payload - A-number verification, B-number status, portability and regulatory checks, a 0-100 risk score with categorical label (Clean / Suspect / High-Risk / Blocked), recommended action and the route-quality and anomaly context behind it - targeting sub-100ms acknowledgement, with bulk batches up to 10,000 numbers. The enterprise API is in build.

How It Works

The process,
step by step

From first peering session to live traffic - a straightforward path with no hidden complexity.

01

API Integration

Connect via REST, WebSocket, SMPP v3.4 or SFTP batch. Pass any E.164 number - or a CDR stream for live route-quality scoring. The enterprise API is in build ahead of the VPX Protocol (Q4 2026).

02

Multi-Source A & B Number Cross-Check

A-number checked against GSMA Call Check, STIR/SHAKEN, Ofcom CLI Authentication and VPX's own CLI history. B-number checked against live HLR, MNP, range allocation and DNO/TPS registries. Source disagreements are themselves a fraud signal.

03

Quality & Anomaly Scoring

The 14-dimension quality engine scores per-route, per-carrier and per-destination across multi-window time bands. A gradient-boosted decision tree surfaces joint-distribution anomalies invisible to single-dimension thresholds - including AIT signatures (bilateral imbalance, uniform-duration, downstream gap).

04

Risk Score & Decision Delivery

Structured JSON response: number status, current MNO, portability history, A-number verification result, 0-100 risk score, categorical label and recommended action - everything needed to route, block or audit in one call.

05

Consensus & Continuous Learning

On the VPX Protocol (Q4 2026), every lookup contributes to a shared intelligence network: Fraud Node blacklist updates are designed to propagate within 60 seconds, and Analytics Node benchmarks commit on-chain hourly via trimmed-median consensus, so every score traces to consensus-verified source data.

Capabilities

Built for
performance

Every capability is designed around real-world wholesale requirements - not a feature checklist.

010-100 + label

Per-Number Risk Score

0-100 fraud probability and a categorical label (Clean / Suspect / High-Risk / Blocked) returned alongside every lookup. Produced by an ensemble model (gradient boosting for interpretable feature scoring + neural network for sequential call patterns) evaluating 47 features per number across call-pattern, destination, network, behavioural and historical categories. On the VPX Protocol (Q4 2026) the score becomes 2/3+ stake-weighted Fraud Node consensus with a 10% per-Operator cap, so no single carrier can flag a number unilaterally. Covers IRSF, Wangiri, CLI spoofing and robocall signatures.

02

AIT & Bilateral Fraud Detection

Artificially Inflated Traffic detection from a composite signal - bilateral imbalance ratio, uniform call-duration (robotic signature) and downstream distribution gap. The cross-network visibility that fully breaks the bilateral-trust model - the fraud category single-carrier systems structurally cannot detect - arrives with the VPX network (Q4 2026). Flagged routes trigger enhanced monitoring and settlement audit, not immediate block.

0314 dimensions

14-Dimension Quality Engine

Per-route, per-carrier and per-destination scoring across 14 quality dimensions - ASR, ACD, PDD, NER, MOS plus jitter, packet loss, SRTT, codec distribution, CLI pass-through, fax success, DTMF reliability, ABR and SER. Multi-window: 1-min, 1-hour, 24h, 7-day, 30-day rolling.

04<2s freshness

Multi-Dimensional Anomaly Detection

Gradient-boosted decision tree over the 14 quality dimensions plus traffic-volume features. Surfaces joint-distribution anomalies invisible to single-dimension thresholds. Webhook alerts in real time with the full feature vector and explanatory drivers attached - no black-box alerts.

05VPX · Q4 2026

Consensus Benchmarks

On the VPX Protocol (Q4 2026): hourly on-chain benchmark commits via trimmed-median consensus (top/bottom 10% excluded), so every quality score, anomaly alert and SLA report will trace to a consensus-verified, immutable source - the auditability foundation that sets VPX apart from proprietary analytics platforms.

06VPX · Q4 2026

On-Chain Evidence Trail

On the VPX Protocol (Q4 2026), every blacklist entry, anomaly alert and benchmark is published on-chain with evidence hashes, full feature vectors and the signing nodes' stake-weighted attestations - so you can ask "why was this blocked?" and get a verifiable cryptographic proof chain, not a black-box answer.

07VPX · Q4 2026

Sub-60s Blacklist Propagation

On the VPX Protocol (Q4 2026), when the Fraud Node network reaches 2/3+ consensus on a malicious number or route, the on-chain blacklist update is designed to propagate to every connected carrier within 60 seconds - versus siloed databases where fraud detected on one network goes unchecked on others for days or weeks.

08Multi-source

A-Number / CLI Validation

Multi-source calling-party verification: GSMA Call Check, STIR/SHAKEN attestation status, Ofcom CLI Authentication policy compliance, ITU-T E.164 conformance, originating-network identity cross-check and VoicePro's own CLI history. Detects spoofing, presentation mismatches and impossible originations no single registry can catch alone.

09200+ routes

B-Number Reachability

Real-time HLR for current MNO, roaming status and IMSI (where permitted), plus MNP across 70+ markets for portability history. Range and prefix validation against ITU-T national plans, premium-rate range flagging and number-allocation registry cross-check. Routes to 200+ countries · 700+ operators.

10

Regulatory Registries

UK Do Not Originate (Ofcom-sourced DNO), Telephone Preference Service (ICO-licensed TPS) and Corporate TPS. International equivalents on supported markets - US DNC, EU national opt-out registries. Real-time and batch modes.

1110K / batch

Bulk List Cleansing

Batch endpoint for pre-campaign list validation - up to 10,000 numbers per request. Returns the full intelligence payload per number: A/B validation, regulatory flags, risk score and recommended action. Ideal for SMS marketing lists, dialler campaigns and CRM data hygiene.

12Zero-risk migration

Dual-Operation Mode

Run Global Insight+ alongside your existing fraud system during a 3-6 month transition. Both checked on every call; fail-safe block if either flags. Compare false-positive and fraud-catch rates side by side, then migrate gradually or operate in dual mode permanently. Removes the cut-over risk that blocks adoption.

Technical Specs

Under the
hood

API ProtocolREST (JSON) · WebSocket stream · SMPP v3.4 · SFTP batch
Response Time<100ms ack · <2s enriched availability · <250ms AI-scored
A-Number SourcesGSMA Call Check · STIR/SHAKEN · Ofcom CLI Auth · ITU-T E.164 · our own CLI history
B-Number SourcesHLR · MNP · range/prefix validation · number-allocation registry · premium-range flags
HLR CoverageRoutes to 200+ countries · 700+ operators
MNP Coverage70+ countries with live porting data
Quality Dimensions14 per route/carrier/destination · 1-min, 1h, 24h, 7d, 30d windows
Fraud Model (per-number)Gradient boosting + neural network · 47 features · Fraud Node stake-weighted consensus on VPX (Q4 2026)
Anomaly Model (traffic patterns)Gradient-boosted decision tree · 14 quality dimensions + traffic features · on-chain consensus on VPX (Q4 2026)
Benchmark Commits (VPX · Q4 2026)Hourly on-chain · trimmed-median (10% / 10% trim)
On-Chain Records (VPX · Q4 2026)Blacklists · anomaly alerts · benchmarks · evidence hashes · node attestations
Data RetentionRaw CDRs: 13 months · Aggregated: indefinite (on-chain from VPX, Q4 2026)
Throughput500 lookups/sec per API key (burst to 2,000) · 100K CDRs/s per Analytics Node (VPX, Q4 2026)
Batch SizeUp to 10,000 numbers per batch request
AuthenticationOAuth 2.0 · API key · IP whitelist
Integration

Plug in anywhere in your stack

Global Insight+ is designed to integrate with your existing switch, gateway, SMS platform or CRM via REST or SMPP, inline with your flow and with no forklift upgrade. The enterprise API is in build ahead of the VPX Protocol (Q4 2026).

REST API (JSON/OAuth2)
WebSocket stream (enriched CDRs + anomaly alerts)
SMPP v3.4 binding
SFTP batch processing
Webhook fraud alerts
On-chain benchmark query (VPX, Q4 2026)
CSV/XML export
Portal dashboard
Asterisk/FreeSWITCH plugins
IP whitelist auth
Who It's For

Built for
your operation

Three distinct profiles that get the most from this product - and why it fits each one.

Carrier · Operator01

Wholesale Carrier

Validate every originating and terminating number before completing a call. Block IRSF, Wangiri and artificial traffic amplification before a fraudulent minute is billed to your network.

Get in touch
SMS · Messaging02

A2P SMS Aggregator

Score every originating sender ID and destination number before routing. Eliminate grey routes, ensure DLR validity and reduce carrier complaints with pre-validated clean traffic submissions.

Get in touch
Contact Centre · Enterprise03

Enterprise Communications

Score inbound numbers before they reach your agents. Block known spam and fraud callers at the network edge. Reduce call centre fraud exposure and protect your brand from CLI spoofing.

Get in touch
Perspective

Trust has a shelf life

A number that was right at sign-up can be wrong today - and the fraud that exploits the gap is already visible at the edge of the network, if you are positioned to see it. This is the shift from holding data to sensing it.

Every fraud check, reset and one-time passcode rests on a quiet assumption: that the number in your records still belongs to the person you think it does. Most of the time it does. But numbers do not sit still - reassigned when customers churn, ported when they switch networks, reactivated months after being disconnected. Whatever you captured at onboarding quietly goes stale, and nobody notices until a message misfires, a real customer is locked out, or a fraud review turns up an assumption that stopped being true.

That is the demand side. The supply side - the fraud itself - is more visible than most realise, if you are standing in the right place. Networks that run SMS firewalls can already see the shape of the attack: the blasts, the grey routes, the bypass traffic, the SIM-swap patterns. The industry now benchmarks it in the open - public dashboards track how many messages national firewalls block per million people per day. The signal exists. The question is who is positioned to read it, and whether they can prove what they saw.

The answer is a shift in emphasis: from gathering information to verifying it, and from verifying at a desk to sensing it at the edge of the network. The GSMA Open Gateway initiative and the CAMARA project are standardising network-derived signals - number verification, SIM-swap detection - so the source of truth can be the network itself, not a form filled in long ago. In parallel, specialists run honeypots: quiet endpoints that attract unsolicited traffic and classify it, because the cheapest place to catch fraud is where it announces itself.

This is the world VoicePro is building for, and it shapes the network we are designing. Two principles guide it. First, verification should be provable, not just returned - the VPX Protocol we are building toward is designed to produce consensus-verified network intelligence, every signal traceable to its source rather than a black box you are asked to trust. Second, sensing should be distributed - fraud-sensing is designed into the network itself, with analytics nodes flagging anomalies that feed a distributed fraud-consensus layer across carriers, rather than a single vendor sensor bolted on afterwards.

There is a human reason this matters to us. VoicePro exists to serve the people telecom has treated as an afterthought - migrant and diaspora communities moving money, minutes and messages across borders, disproportionately targeted by number-based fraud precisely because the verification layer beneath them is weakest. Provable, current verification and network-native sensing are not back-office plumbing for them; they decide whether a remittance reaches family or a scammer. Trust remains essential - but it has a shelf life. The organisations that come through the next few years well will not be the ones holding the most data. They will be the ones close enough to the network to see what has changed, and able to prove it, at the moment it counts.

VoicePro Plus is a UK telecom and fintech company. Our UK mobile service launches on 1 November 2026; the VPX Protocol mainnet is targeted for Q4 2026 and is in active design.

Start a Discussion

Ready to activate Global Insight+?

Our solutions team responds within four business hours. Committed volume agreements, per-second billing and dedicated account management available from day one.

Get in Touch