Leave 32% less personal data exposed than AWS Comprehend, 79% cheaper

Hide names, account numbers, addresses and customer IDs in chats, tickets and emails before the text reaches an LLM, a log or a vendor. SIE, an open-source inference engine, runs two open models, GLiNER PII and NuNER Zero, behind one API call.

Get started Run in your cloud

SIE GLiNER PII and NuNER Zero

All personal data masked

Payment confirmation

Transaction ID: ID
Date: 08/12/2022
Time: 14:36:51 GMT

Payer:
Name: name
Customer ID: customer ID

Payee:
Name: name
Address: address

Transaction Amount: £150.99

Personal-data detectors leave IDs and names readable

AWS Comprehend

Leaves the employee ID and a name readable

Support chat log

| date time date time | Employee ID: O4878342 | Sure, Maya. I can see you're our valued customer with the employee ID O4878342. Now, to track your order, I just need your order number. |

OpenAI Privacy Filter

Leaves the customer's name readable

Subject: Exclusive Offers Just For You, Hugh!

Dear Hugh,

We hope this email finds you well. We're excited to bring you some exclusive offers tailored just for you!

Malwarebytes An AI chat app left 300 million messages from 25 million users in an open database
The researcher claims to have accessed 300 million messages from over 25 million users due to an exposed database.
9 February 2026
Security Magazine 3.7 million records from Sears' AI assistant, names and addresses included, sat open online
Cybersecurity Researcher Jeremiah Fowler discovered three different publicly exposed databases containing 3.7 million records, of which all observed data appeared to belong to Sears Home Service.
23 March 2026

Masks more than AWS Comprehend, 22x faster than GPT-6 Luna

Run this eval

Personal data masked

660 documents ↑

GPT-6.1 Sol

GPT-6.1 Sol 95.7%

GPT-6 Luna (high reasoning)

GPT-6 Luna (high reasoning) 94.7%

Claude Haiku 5.5

Claude Haiku 5.5 94.4%

SIE GLiNER PII and NuNER Zero

SIE GLiNER PII and NuNER Zero 90.9%

AWS Comprehend

AWS Comprehend 84.2%

OpenAI Privacy Filter

OpenAI Privacy Filter 69.7%

Time per document

median ↓

SIE GLiNER PII and NuNER Zero

61 ms SIE GLiNER PII and NuNER Zero

GPT-6 Luna (no reasoning)

GPT-6 Luna (no reasoning) 1.39 s

Cost per 1M documents

USD ↓

OpenAI Privacy Filter

$23 OpenAI Privacy Filter

SIE GLiNER PII and NuNER Zero

$73 SIE GLiNER PII and NuNER Zero

Claude Haiku 5.5

$188 Claude Haiku 5.5

AWS Comprehend

$351 AWS Comprehend

GPT-6 Luna (high reasoning)

$352 GPT-6 Luna (high reasoning)

GPT-6.1 Sol

GPT-6.1 Sol $2,487

Tested: 660 synthetic financial documents, 1,792 pieces of personal data
Scored: A piece counts when every letter and digit is hidden; punctuation may stay readable
Prices: LLM, AWS and Privacy Filter GPU list prices; SIE's Serverless rate
Timing: SIE timed on a GPU server with the client alongside; providers over their public APIs

Send text, get every span to mask

from sie_sdk import SIEClient
client = SIEClient(api_key="sk-sie-…")
result = client.extract(
    "modelsuperlinked/pii-redact",
    {"text": "text{"date_time": "1978-03-29 17:28:13", "name": "Emma A. Platini", "street_address": "19325 John Motorway, Apt. 488"}
{"date_time": "2022-01-02 14:35:18", "name": "John D. Smith", "street_address": "456 Oakwood Avenue, Unit 3C"}
{"date_time": "2021-12-15 09:19:21", "name": "Alex N. Thompson", "street_address": "334 King's Cross, Flat 15B"}"},
)
for span in result["entities"]:
    print(span["label"], span["text"])

import { SIEClient } from '@superlinked/sie-sdk';

const client = new SIEClient(undefined, { apiKey: 'sk-sie-…' });

const result = await client.extract(
  'superlinked/pii-redact',
  { text: "{\"date_time\": \"1978-03-29 17:28:13\", \"name\": \"Emma A. Platini\", \"street_address\": \"19325 John Motorway, Apt. 488\"}\n{\"date_time\": \"2022-01-02 14:35:18\", \"name\": \"John D. Smith\", \"street_address\": \"456 Oakwood Avenue, Unit 3C\"}\n{\"date_time\": \"2021-12-15 09:19:21\", \"name\": \"Alex N. Thompson\", \"street_address\": \"334 King's Cross, Flat 15B\"}" },
  { labels: [] }, // the model's default types
);
console.log(result.entities);

curl https://api.superlinked.com/v1/extract/superlinked%2Fpii-redact \
  -H "Authorization: Bearer sk-sie-…" \
  -H "Content-Type: application/json" \
  -d "{\"items\":[{\"text\":\"{\\\"date_time\\\": \\\"1978-03-29 17:28:13\\\", \\\"name\\\": \\\"Emma A. Platini\\\", \\\"street_address\\\": \\\"19325 John Motorway, Apt. 488\\\"}\\n{\\\"date_time\\\": \\\"2022-01-02 14:35:18\\\", \\\"name\\\": \\\"John D. Smith\\\", \\\"street_address\\\": \\\"456 Oakwood Avenue, Unit 3C\\\"}\\n{\\\"date_time\\\": \\\"2021-12-15 09:19:21\\\", \\\"name\\\": \\\"Alex N. Thompson\\\", \\\"street_address\\\": \\\"334 King's Cross, Flat 15B\\\"}\"}]}"

Build the "Find personal data to mask" capability into my app using the Superlinked Inference Engine (SIE).

Context
- SIE is an OpenAI-style inference API. Python SDK: `from sie_sdk import SIEClient`; TypeScript: `@superlinked/sie-sdk`.
- Base URL: https://api.superlinked.com (or my regional endpoint). Auth: Bearer key from env `SIE_API_KEY` (never hard-code it).
- Model: superlinked/pii-redact (SIE primitive: /extract). Keep the model id configurable.

Task
- Input: a block of text.
- Behaviour: return the spans of personal data found, each with its type and character offsets; my code masks them.
- Call the selected SIE primitive once per request and map the response into your domain type.

Deliverables
- A typed client wrapper, an application-level function for this task, error handling for timeouts/empty input, and unit tests with a stubbed client.
- Wire it into my existing stack (ask me which framework if unclear) and add a short usage example.

Run

Output

entities · 6start–end

person Emma A. Platini 46–61
street address 19325 John Motorway, Apt. 488 83–112
person John D. Smith 161–174
street address 456 Oakwood Avenue, Unit 3C 196–223
person Alex N. Thompson 272–288
street address 334 King's Cross, Flat 15B 310–336

masked · 6

{"date_time": "1978-03-29 17:28:13", "name": "name", "street_address": "address"}
{"date_time": "2022-01-02 14:35:18", "name": "name", "street_address": "address"}
{"date_time": "2021-12-15 09:19:21", "name": "name", "street_address": "address"}

Mask personal data 79% cheaper than AWS Comprehend

Serverless API

We host the models; you call one API.

No idle GPUs, pay for what you use
Fits your stack: SDK, API, CLI, MCP
Self-host the same open-source engine
SOC 2 Type 2, US or EU data residency
Create an account to try Serverless

Get started

Self-hosted (Kubernetes)

Your data stays in your cloud, backed by our engineers.

Terraform modules for your cloud
Apache-2.0, same engine as Serverless
GPU workers scale to zero between jobs
Separate worker pools for each tenant

Agent prompt Deploy SIE to our AWS account with the superlinked/sie/aws Terraform module. Docs: superlinked.com/docs/deployment.

Deploy SIE to our Google Cloud project with the superlinked/sie/google Terraform module. Docs: superlinked.com/docs/deployment.

Deploy SIE to our Azure subscription with the superlinked/sie/azure Terraform module. Docs: superlinked.com/docs/deployment.

Deploy SIE to our Alibaba Cloud account with the superlinked/sie/alicloud Terraform module. Docs: superlinked.com/docs/deployment.

Talk to us Deploy guide

Run locally

The open-source server on your own machine.

Runs on your NVIDIA GPU
One container, no cluster
Process your data fully offline
Same SDK and model IDs as Serverless

docker run --gpus all -p 8080:8080 ghcr.io/superlinked/sie-server:latest-cuda12-default

Quickstart

Switch to SIE in 5 minutes

import boto3

comprehend = boto3.client("comprehend")


def find(text):
    found = comprehend.detect_pii_entities(
        Text=text, LanguageCode="en"
    )
    return [
        (e["BeginOffset"], e["EndOffset"])
        for e in found["Entities"]
    ]

from presidio_analyzer import AnalyzerEngine

analyzer = AnalyzerEngine()


def find(text):
    found = analyzer.analyze(text=text, language="en")
    return [(e.start, e.end) for e in found]

SIE

from sie_sdk import SIEClient

client = SIEClient(api_key="sk-sie-…")


def find(text):
    found = client.extract(
        "superlinked/⁠pii‑redact",
        {"text": text},
    )
    return [(e["start"], e["end"]) for e in found["entities"]]

from sie_sdk import SIEClient

client = SIEClient(api_key="sk-sie-…")


def find(text):
    found = client.extract(
        "superlinked/⁠pii‑redact",
        {"text": text},
    )
    return [(e["start"], e["end"]) for e in found["entities"]]