Skip to content
A secure knowledge workspace for people and AI agents.

One knowledge.
Your permissions, everywhere.

Used more. Leaked never.

Documents, structured data and the answers AI gives: everything your company knows, in one environment. People and AI agents see only what they are allowed to. Personal data is masked everywhere by a policy you set once, and every action is left as an approval and a history entry.

Start ModooKM for freeEvidence · 15-minute demoTalk about Analytics

The screen shown here is real. Run the same query as two users with different permissions and different values get masked. 15-minute demo scenario →(opens in a new window)

Assistant screen with the customer-grade policy document set as context, asked how many customers need regrading against the VIP rule. The answer is 256, and in the table below customer names and phone numbers are masked.The same query run by the account named as the policy exception. Names and phone numbers appear in full in the same table.
The assistant reads the grade rule from a document, checks 500 customers in the database and answers 256. Names and phone numbers are masked by permission. Real screens from a demo tenant with synthetic customer data.
Where are you coming from

One company, three services. Go straight to the one you need.

This page keeps it short. Pick your situation and the service page takes it from there.

ModooKM · Individuals · teams of 2–20
We have plenty of documents, but when someone asks, nobody can answer on the spot

ModooKM

Upload PDF, Office and HWP files. Ask a question and the answer comes back with the file name and page. Even the administrator cannot open your private space.

Free · 1M tokens per month · 1 GB storage · first answer within 5 minutes of signing up
Start for free
PAASUP Evidence · Public sector · finance · manufacturing/defense · ISO/ISMS
What AI did leaves no trace. At audit time we cannot answer "who approved this"

Evidence

A work system where even AI actions become evidence. Document edits, approvals, permission changes and AI proposals are recorded in an immutable history and audit log, with who did what and when.

15-minute demo → 2-week PoC (with one department's documents)
Book a 15-minute demo
PAASUP Analytics · Data teams · partners
Documents here, databases there, masking somewhere else. One question means three systems

Analytics

Documents and data in a single question. Whatever must be masked stays masked wherever the query runs. When you outgrow it, DIP takes over.

Consultation → PoC on one data source
Request a consultation
How it fits together

Services on top, platform underneath.
Permissions and policy are the same everywhere.

The three services run on the same repository. When scale grows, the platform below takes the processing. Who can see what does not change.

Service layer · what people use, on the knowledge core
ModooKM
Gather. Where personal documents first become knowledge
ModooKM · modookm.com
Evidence
Record. What people and AI did, kept as evidence
PAASUP Evidence
Analytics
Ask. One question across documents and data
PAASUP Analytics
One permission rule, one de-identification policy, approvals and immutable history. All of it carries down as is
Platform layer · the infrastructure services stand on, run by IT
PAASUP DIP · Data Intelligence Platform
Batch, virtualization, streaming and LLM serving on Kubernetes in one catalog. Runs in an air-gapped network at public-SaaS quality.
GS Certification Grade 1 · registered patents · ISO/IEC 27001 (scope: design, development, supply, deployment and maintenance of DIP for air-gapped environments) · 400M records a day for Bucheon ITS
A document one person uploads becomes team knowledge, and team records become organizational evidence. When document terms link to data tables, nothing migrates. You turn a feature on.
Why nothing leaks

Unified, protected, working together.

The three screens below reproduce the same flow in a demo tenant with a synthetic organization and synthetic HR data.

Screen with the Customer Grade Policy document open and the assistant asked, in reference mode, for the GOLD criterion and the current number of GOLD customers. The grade table sits on the left; the quoted criterion and the count from the customers table on the right.
The GOLD criterion quoted from the document and the customer count from the database in one answer · demo tenant
Unified

Ask and analyze in one environment

Documents, unstructured and structured data live in one repository, one catalog, one question box. Document terms meet data tables, and one question gets an answer from documents and the database together.

The content filter binding list. Document paths and data paths carry redact_pii and mask_partial policies, with one named exception subject.
One de-identification policy over document paths and data paths · demo tenant
Protected

One permission, one policy, no exceptions

People, AI and shared links all pass through the same permission gate, whether the path is viewing, search, SQL or ETL. Personal data is masked at read time by a policy set once. The original is never touched.

Built on 2 pending and 3 registered patents

A document change approval screen. The drafter and the final approver sit on the approval line and the final approval is still pending.
Where an agent proposal passes through approval · demo tenant
Together

AI proposes, people approve

Agents cannot write directly. Proposals go through approval before they land, and what people did and what AI did share one history. That is why you can hand more to AI.

We are our first customer

PAASUP runs on it.

Customer logos will appear here as they are secured. Until then, PAASUP itself is the evidence.

PAASUP's policies, contracts, meeting notes, approvals and security operations records all live in this system. Our ISO/IEC 27001 audit evidence comes from here too. The screens below reproduce the same flow in a demo tenant with a synthetic organization and synthetic HR data.

Operating records · PAASUP tenantLast 7 days
Policy revisionPolicy change → approval → approved version history. Who approved what and when, in an immutable recordv1.3 · approved
Permission grantsWhich document scopes and data scopes each person has been granted, answered on screengranted 2
AI proposalA document draft written by an agent, approved and committedREQ-581
Audit evidenceDocument history and audit logs submitted for the ISO/IEC 27001 surveillance auditGI-260171
The data source tab of the permission screen. An HR data scope lists the tables it reaches and the people it is assigned to.
Permission grants · which scope went to whom, answered on screen
The audit log screen. Sign-in, department change and configuration change events are listed with actor, time and before and after values.
Audit log · the combined event history used as ISO/IEC 27001 A.8.15 evidence
Certifications · partners · track record

What the company holds.

Certification scopes are quoted as written. Partner and customer logos are used with confirmed permission.

GS Certification Grade 1PAASUP DIP · TTA software quality certification
3 patents registered · 2 pendingRegistered: cross-network binary repository (Binup) · organizational knowledge management · integrated open-source SaaS management (PCT · US filing) / Pending: unified access control across heterogeneous resources · read-time de-identification
ISO/IEC 27001:2022GI-260171 · scope: design, development, supply, deployment and maintenance of DIP for air-gapped environments
Venture BusinessCertified company
What we believe

What a company knows is scattered across documents, databases and people's heads.

AI offers to gather it all and answer. Yet who may see what is decided differently by every tool.

We reverse the order. First we set one permission model, then we gather knowledge on top of it.

We do not separate documents from data. We do not separate people from AI. The only thing we separate is permission.

AI proposes and people approve. Whoever acted, it lands in the same history.

So the company's knowledge is used more and never leaves its permissions.

Q & A

FAQ

PAASUP builds a secure knowledge workspace for people and AI agents. It offers three services (ModooKM · PAASUP Evidence · PAASUP Analytics) and one platform (PAASUP DIP).

01What does PAASUP do?

PAASUP lets a company use everything it knows, documents, structured data and AI answers, in one environment. The service layer has ModooKM, PAASUP Evidence and PAASUP Analytics. The platform layer has PAASUP DIP.

02How do the three services differ?

ModooKM is an AI knowledge base for individuals and small teams. PAASUP Evidence is an audit-ready work system for organizations where even what AI did is kept as evidence. PAASUP Analytics is an agentic analysis service that asks documents and data in one question. All three run on the same repository and the same permission rules.

03Do AI agents follow permissions too?

Yes. People, AI agents and shared links all pass through the same permission gate, whether the path is viewing, search, SQL or ETL. Personal data is masked at read time by a policy set once, and the original is never touched.

04What certifications and track record does PAASUP have?

GS Certification Grade 1 for PAASUP DIP, 3 registered patents and 2 pending, ISO/IEC 27001:2022 (GI-260171, scope: design, development, supply, deployment and maintenance of DIP for air-gapped environments), Venture Business certification, NVIDIA Inception and SUSE One partnerships, and 400 million records a day processed for Bucheon ITS.

05Where should I start?

Individuals and small teams can start ModooKM for free. Organizations can book a 15-minute Evidence demo or request an Analytics consultation through the contact form.

Contact

Where shall we start.

The 15-minute demo is online. The 2-week PoC starts with one department's documents.

ModooKM

ModooKM

Free · 1M tokens per month · 1 GB storage · first answer within 5 minutes of signing up

Consultation
PAASUP Evidence

PAASUP Evidence

15-minute demo → 2-week PoC (with one department's documents)

15-minute demo
PAASUP Analytics

PAASUP Analytics

Consultation → PoC on one data source

Consultation