AI-Native PII Tokenization Gateway — Banking Analytics Preserved

Your LLMs. Zero PII.
Full Analytics.

The drop-in AI privacy gateway that tokenizes sensitive data before any prompt leaves your organization — then reverses it on the way back. Role-aware. Audit-logged. Built for regulated finance.

Prevent GLBA violations · Satisfy GDPR · Preserve CCPA rights · Deploy in under 30 minutes

vaultex — tokenization gateway — live
OUTBOUND (raw prompt — your network)
Summarize loan for Jane Smith — SSN 123-45-6789, account ACC-00198234, email jane@acme.com. Balance $42,500, credit score 742, risk LOW.
VAULTEX GATEWAY — <8ms tokenization
INBOUND TO LLM (safe — external API)
Summarise loan for {{PERSON_1}} — SSN {{SSN_1}}, account {{ACCT_1}}, email {{EMAIL_1}}. Balance $42,500, credit score 742, risk LOW.

Product

API proxy or chatbot?

Both. Vaultex is a gateway that fits into your existing workflow — no matter how you work with LLMs.

Drop-in
REST API Proxy

One line change. Your existing SDK.

Point any Anthropic, OpenAI, or Ollama SDK at the Vaultex gateway URL. Every outbound prompt is auto-tokenized. Every response is auto-detokenized. Zero refactoring.

# Before
client = anthropic.Anthropic(api_key="sk-ant-...")

# After — one change
client = anthropic.Anthropic(
  base_url="https://your-vaultex-host:8000",
  api_key="sk-ant-..."
)
No code
Chat UI

Browser-based. Upload CSV. Ask questions.

Open the built-in web interface at /chat. Upload a CSV of customer records, start a conversation, and watch PII get masked in real time — role-filtered per your RBAC config.

Analytics
CSV Analytics Mode

Upload data. Query at scale. PII stays masked.

Drop a CSV with thousands of loan records into the /chat interface. Ask portfolio-level questions. The LLM sees tokens, computes real analytics, and you see role-filtered answers.

GLBA Aligned
GDPR Ready
CCPA Aware
Microsoft Presidio NER
Anthropic | OpenAI | Ollama
Append-Only Audit Trail
<8ms
Added latency per request
14+
PII entity types detected
100%
Prompts tokenized before LLM
0
PII stored in plaintext

The Problem

AI and PII are on a collision course

Financial institutions face compounding pressure to adopt AI — but every unguarded prompt is a potential compliance incident.

Analysts paste SSNs into ChatGPT

Every time a team member copies customer data into an AI prompt, your org risks a GLBA violation with no log of what left your network or which model processed it.

No audit trail for regulators

Cloud LLMs process and may retain prompt data. Without a proxy layer, you have zero regulator-ready evidence of what sensitive information reached which external model.

Junior staff see VP-level data

Most AI tools have no row-level access control. A junior analyst sees the exact same AI response as a VP — including full PII they should never be authorized to access.

Capabilities

Enterprise-grade protection, out of the box

Everything your team needs to safely connect to any LLM without exposing customer data, regulated PII, or financial records.

Presidio NER Detection

Microsoft Presidio NER catches PERSON, SSN, ACCOUNT_NUMBER, LOAN_ID, EMAIL, PHONE, and DATE_OF_BIRTH before any prompt crosses your network boundary.

Reversible Tokenization

PII becomes {{PERSON_1}} — not random noise. The same person gets the same token in every message. Responses auto-detokenize on the way back.

Analytics Preserved

Balances, credit scores, rates, and risk flags are NEVER masked. Your LLM can still compute averages, flag high-risk loans, and run full distributions.

Multi-LLM Routing — PII-stripped

Route to Anthropic Claude, OpenAI GPT-4, or self-hosted Ollama. The gateway tokenizes PII locally before any request reaches a cloud API — so neither Anthropic nor OpenAI ever sees a real name, SSN, or account number.

Role-Based Decryption

Junior analysts see {{PERSON_1}}. Senior analysts see the name. VPs see everything. RBAC enforced at the token level, not the application layer.

Append-Only Audit Trail

Every request, PII detection event, and auth failure is logged with correlation IDs. 30-day retention. Regulator-ready admin console included.

View full security architecture

Integration

One line of config.
Everything else is automatic.

Vaultex is a drop-in proxy. Point your existing SDK at our gateway endpoint and every prompt is automatically tokenized — no code changes, no schema migrations, no training required.

Works with every Anthropic, OpenAI, and Ollama SDK
Processes any language or framework transparently
Deployable in Docker in under 30 minutes
REST API for custom integrations
integration.py
# Before Vaultex
import anthropic
client = anthropic.Anthropic(
    api_key="sk-ant-..."
)

# After Vaultex — one change
client = anthropic.Anthropic(
    base_url="https://your-gateway-host:8000",
    api_key="sk-ant-..."
)

# Every prompt is now auto-tokenized.
# Zero code changes beyond base_url.

Banking Analytics

Regulators want privacy.
Analysts want numbers.
You get both.

Vaultex only masks direct personal identifiers. Financial dimensions — the data your analysts actually need — flow through untouched. Your LLM can compute averages, distributions, and risk clusters without ever seeing a real name or SSN.

Average credit score by state: computed from real values
High-risk loan segmentation: based on actual DPD and risk flags
Interest rate distribution: exact rates, no tokens
Portfolio balance totals: real dollar figures, preserved
FIELD CLASSIFICATION — Jane Smith record
Balance
$42,500.00REAL
Credit Score
742REAL
Interest Rate
4.75%REAL
Days Past Due
0REAL
Risk Flag
LOWREAL
Loan Type
MortgageREAL
State
CAREAL
Name
{{PERSON_1}}MASKED
SSN
{{SSN_1}}MASKED
Email
{{EMAIL_1}}MASKED
Account #
{{ACCT_1}}MASKED
Phone
{{PHONE_1}}MASKED

How It Works

Zero to protected in 3 steps

01

Connect Your LLM

Choose Anthropic, OpenAI, or a local Ollama model. Enter your API key — encrypted at rest with Fernet. Point your SDK at the Vaultex base URL. Done in 30 seconds.

02

Chat With Your Data

Upload a CSV with customer records. The gateway intercepts every prompt, tokenizes PII, and forwards the sanitized version to your chosen LLM. Analytics fields pass through intact.

03

See Role-Filtered Results

Switch personas live. Watch how a junior analyst sees tokens while a VP sees real names. RBAC enforced at the token layer — no application code changes needed.

Why Vaultex

Purpose-built for banking. Nothing else comes close.

Generic DLP tools scan files. Vaultex tokenizes live AI prompts — and preserves analytics. That is a different product entirely.

FeatureVaultexDLP ToolsDirect APIManual
Reversible tokenization
Analytics fields preserved~
Role-level PII decryption
Append-only audit trail
On-premise / self-hosted
Multi-LLM routing~~
Zero code changes required
GLBA evidence pack~~

~ = partial / varies by vendor implementation

Access Control

Who sees what

Role-Based Decryption — enforced at the token layer, not the application layer.

RoleReal PIIEntities Visible
Junior AnalystNone — all tokens shown
Senior AnalystPERSON, EMAIL
VP RiskAll personal entities
AdminFull PII + Audit Console
Product Walkthrough

Every prompt. Zero PII leakage.

Watch a live banking risk query flow through the Vaultex gateway — tokenized on the way out, detokenized on the way back, fully audit-logged throughout.

vaultex — privacy gateway⚠ PII Detected
Step 1 / 4Raw PromptApp sends prompt with PII
your-app → vaultex-gateway :8000
POST /v1/chat/completions HTTP/1.1
Authorization: Bearer sk-ant-...
{ "messages": [{ "role": "user", "content":
"Analyse risk for Jane Smith (SSN: 123-45-6789,
email: jane.smith@acme.com, phone: 415-555-0192).
Account ACC-00198234, balance $42,500.
Credit score 742, Risk Flag: LOW." }]
presidio NER — entities detected
Scanning with en_core_web_lg…
● PERSON "Jane Smith"
● US_SSN "123-45-6789"
● EMAIL "jane.smith@acme.com"
● PHONE "415-555-0192"
4 entities found · 0 false-positives
1 / 4

Simulated walkthrough · Real gateway uses Microsoft Presidio NER with 14+ entity types · avg +12ms overhead

Try It Now

See tokenization in action

Edit the sample text, then click Tokenize. PII is detected client-side in this preview — the real gateway uses Microsoft Presidio NER with 14+ entity types.

vaultex — browser preview
INPUT (edit me)

Pricing

Start free. Scale with confidence.

The core tokenization engine is MIT open-source and always will be.

We earn by selling the features regulated institutions actually need: cloud LLM routing, SOC 2 reports, GLBA evidence packs, SSO, extended retention, and dedicated SLAs. Open source builds trust; enterprise features fund the team.

Starter
Free

Self-host with Ollama. Full tokenization engine, no credit card, no time limit.

Ollama (local models) only
Up to 3 users
Core PII tokenization (6 entity types)
7-day audit log retention
Community support
MIT open-source license
Most Popular
Professional
$299/mo

Unlock Anthropic + OpenAI routing, full compliance tooling, and team access.

Anthropic Claude + OpenAI GPT + Ollama
Up to 25 users
Full PII entity coverage (14 types)
90-day audit retention
CSV analytics mode
Role-based export controls
Priority email support
Enterprise
Custom

For regulated institutions — on-prem VPC, SOC 2, GLBA evidence packs, and SLAs.

Unlimited users
On-prem / private VPC deploy
SOC 2 Type II report
Custom RBAC policies
GLBA / GDPR evidence pack
Dedicated Slack support channel
99.9% uptime SLA

All plans include end-to-end encryption, RBAC, and audit logging. Compare all features in detail

FAQ

Questions we get asked

Read our full compliance documentation
Early Access — 3 months free at launch

Early Access

Get notified when
Professional launches

Be first to access multi-LLM routing, 90-day audit retention, and role-based export controls. Early access members receive 3 months free.

Opens your email client. No spam. One update at launch.