Skip to content
Self-hosted enterprise AI platform · v1.5.0 · 2026-09

The platform for running AI
in regulated environments

Agents, data, evaluation, and governance on one platform. Cloosphere calls LLMs directly from your own account, and conversations, documents, and logs stay in your environment. None of it passes through Cloocus servers.

Deploy to your cloud, on-premises, or an air-gapped network DEPLOY
Audit logs and guardrails included, no add-ons GOVERN
LLM cost with no resale markup COST
TRACE model call → web search → knowledge base search → DB schema lookup → SQL execution → chart generation 17.66s · 42,780 tokens
Agent execution trace
An agent execution trace. For a single question, the model call, web search, knowledge base search, DB schema lookup, SQL execution, and chart generation are all recorded in order, with how long each step took.
5+ LLM connection types OpenAI · Azure OpenAI · Azure AI Foundry · Vertex AI · Ollama + OpenAI-compatible
8 Databases you can query in plain language PostgreSQL · MySQL / MariaDB and 6 more
14+ Document formats Including HWP · 50+ code file types
5 Vector DBs · search indexes pgvector · Elasticsearch · OpenSearch and 2 more
3 Clouds + on-premises Azure · Google Cloud · AWS
Monthly Release cadence New models and specs added with each release
LLM
  • OpenAI
  • Azure OpenAI
  • Azure AI Foundry
  • Vertex AI
  • Ollama
  • OpenAI-compatible
DB
  • PostgreSQL
  • MySQL / MariaDB
  • Microsoft SQL Server
  • Oracle
  • Snowflake
  • Databricks
  • Google BigQuery
  • Azure Synapse
Cloud
  • Azure
  • Google Cloud
  • AWS
Credentials of Cloocus, the company behind Cloosphere Microsoft Azure Expert MSPOpenAI Select Partner12 Azure Advanced SpecializationsMicrosoft Partner of the Year Global Finalist
What you can do

Four pillars: agents, data, evaluation, and governance

To run AI as part of daily operations, an organization needs all four on the same platform.

Agents & flows

Combine tools, knowledge, databases, and glossaries to answer layered questions

Chat across several LLMs in one view, build flows in a no-code visual builder, generate PPT, Word, and Excel files on your company templates, and connect MCP, OpenAPI, Google Workspace, and Microsoft Teams tools — all in one place.

More about Agents & flows →
AGENT Agents & flows
Combine tools, knowledge, databases, and glossaries to answer layered questions
Data

Turn plain questions into SQL and ask your 8 databases directly

A knowledge base that handles 14+ document formats including HWP, a table-relationship panel that infers joins from foreign keys and naming rules, a SQL editor that puts write queries through approval, and a glossary that carries your in-house terms into answers.

See data in detail →
DATA Data
Turn plain questions into SQL and ask your 8 databases directly
Evaluation · observability

Record the whole execution path, from model call to guardrail

Traces in tree, timeline, and token views; automatic scoring of retrieval quality, faithfulness, and answer quality from 0 to 1; Arena comparisons with per-model Elo; and usage and cost estimates by model, user, organization, and agent.

See evaluation · observability in detail →
USAGE Evaluation · observability
Record the whole execution path, from model call to guardrail
Governance

Guardrails, audit, permissions, and the Code Gateway are part of the license

Audit logs and guardrails are included from the Standard tier, and access controls are available on every tier. There is no separate add-on just for controls.

12 PII types detected · 19 audited resource types · 4 permission tiers

See governance in detail →
AUDIT Governance
Guardrails, audit, permissions, and the Code Gateway are part of the license

Works with the clouds, databases, and documents you already use

LLM Every model available in your account
OpenAI Azure OpenAI Azure AI Foundry Vertex AI Gemini · Claude Ollama OpenAI-compatible Private gateways · vLLM, etc.
Databases Natural language → SQL
PostgreSQL MySQL / MariaDB Microsoft SQL Server Oracle Snowflake Databricks Google BigQuery Azure Synapse
Cloud Deployed into your account
Azure Google Cloud AWS
Vector · search indexes
pgvector Elasticsearch OpenSearch Azure AI Search Vertex AI Search
Document formats
PDF DOCX PPT · PPTX XLS · XLSX CSV HWP · HWPX Markdown HTML XML RST EPUB EML · MSG TXT JSON 50+ code file types
Extraction · OCR · web search
Apache Tika Docling Azure Document Intelligence Google Document AI Mistral OCR LLM Vision 16 web search providers
Authentication
Microsoft Entra ID Org sync Google GitHub OIDC LDAP

Version-by-version support scope and limitations are published as a table on the integration specs page. See all integration specs →

All logos and names are trademarks of their respective owners. They indicate supported integrations and do not imply any partnership or endorsement. Full trademark and license notices

Deployment

Deployed wherever your data lives

Your cloud

Deployed into your account on Azure · Google Cloud · AWS. Operations are handled by Cloocus as an MSP, or by your own team.

  • LLMs called directly from your account
  • Entra ID · Google · OIDC SSO
  • Azure Key Vault envelope encryption (KMS)

Air-gapped network

A signed license file is verified locally inside the deployment, so there is no communication with a license server. LLMs run as internally hosted models or through an approved private endpoint.

  • Ollama · private OpenAI-compatible endpoints
  • Runs with web search and external tools disabled
  • Includes a pre-deployment readiness check

On-premises Coming soon

Deployed into your own infrastructure. Connects directly to the vector DB and LDAP you already run.

  • Logo, favicon, and splash screen can be replaced (white label)

3 white-label assets (logo · favicon · splash screen) · 4 embed widget + SSO options · 55 UI languages

Three options — what actually differs

ItemCloosphereSelf-service SaaSSelf-run open source
Controls Audit logs (19 resource types), guardrails, and access controls included in the license Audit and guardrails are a paid add-on or limited in scope Build and maintain it yourself
LLM cost Your account's official pricing, no platform margin Vendor-set pricing (may differ from the official rate) Contract directly
Data location Your cloud, on-premises, or an air-gapped network The vendor's cloud Your own infrastructure
Deployment · support Project-based deployment + MSP operational support + monthly releases Configuration within the contract scope Your own staff
Observability Traces, automatic quality evaluation, and usage cost estimation built in Usage reports Integrate a separate tool

How different organization types use it

Financial services

  • Air-gapped deployment
  • Audit logs with before/after values
  • 6 Korea-specific PII detection types
  • Approved model list
Discuss deployment scope →

Public sector

  • Per-department token limits
  • Org → department → group permissions
  • 6 kinds of activity history
  • On-premises deployment
Discuss deployment scope →

Manufacturing · large enterprise

  • Natural language analysis of internal databases
  • Multi-org · Entra ID sync
  • White-labeled internal service
  • Google Workspace · Teams integration
Discuss deployment scope →

Development organizations

  • Control over Claude Code · Codex · Gemini CLI · Cursor
  • Code Gateway guardrails
  • Call logs and block reasons
  • MCP · OpenAPI tool approval
Discuss deployment scope →

Request a demo →

What changed in each release

1.5.0 2026-09

Only verified question–SQL examples saved automatically, MCP tool exposure mode selection, GPT-6 family support

1.4.0 2026-08

Workspace activity history, selection rationale for multi-database connections, Code Gateway guardrail redesign, chat stop/reconnect and permission/session hardening

1.3.0 2026-07

Knowledge Base library, full flow builder redesign, usage visualization and cost estimation, Google Docs/Sheets/Slides creation and editing, workspace skills

Earlier: 1.2.0 (June 2026) Google Workspace integration, ERD, prompt version control; 1.1.0 (May 2026) SQL editor, clarifying questions and approval, document generation; 1.0.3 (April 2026) Enterprise KMS, MCP OAuth

All release notes →

Common questions before adopting Cloosphere

How does pricing work?

It's a license based on a build fee plus an annual operations contract, and what's included depends on the tier — 4 tiers (Basic · Standard · Professional · Enterprise). The full feature-by-tier table is published on the pricing page; pricing itself is quoted based on your configuration. LLM usage is billed directly to your own cloud account.

Who sets up the LLM account?

You connect your own account (Azure OpenAI, Vertex AI, etc.). If you don't have one, Cloocus can help set it up and run it as an MSP.

Where is data stored?

Conversations, documents, and logs are stored only in the environment you deployed to (your cloud account or on-premises) and are never sent to Cloocus servers. Your questions are only sent to the LLM account you connected, and secrets can be kept using envelope encryption.

Does it work in an air-gapped network?

The license verifies a signed file locally inside the deployment, so it never talks to a license server. LLMs run as internally hosted models (like Ollama) or through an approved private endpoint. Outbound features like web search are disabled, and a pre-deployment readiness check identifies any external dependencies.

How long does deployment take?

It's scoped as a project based on air-gap requirements, SSO, and database integration. We provide a scope-based timeline when you request a demo.

Request a demo

See how it runs in your environment

We set up a demo environment based on your deployment method, integration scope, and governance requirements.

What happens after you reach out

  1. 01
    Reply

    A member of our team will review your request and contact you.

  2. 02
    Requirements call

    A roughly 30-minute call to confirm your deployment environment, integration targets, and governance requirements.

  3. 03
    Demo

    We demonstrate in a demo environment set up to match what we confirmed.

  4. 04
    Proposal

    We put together a deployment scope, timeline, and quote for you.

For urgent inquiries: +82 2-597-3400 · marketing@cloocus.com

Areas of interest (select all that apply)
Organization size

Just need the brochure? → Get it by email