Self-hostedGDPR-nativeOpen Source Core

An AI Operating System
that remembers.

Live in 5 minutesDocker Compose. One command. Done.
No configurationJust one API key. Everything else has sensible defaults.
Fits your setupREST API, MCP Server, CLI. Whatever you already use, works.
7-Layer MemoryGlobal Workspace ReasoningPersistent AgentsFSRS Spaced RepetitionKnowledge GraphVoice & Vision55+ ToolsCLI Agent
9,228Tests
6Pillars
< 5 minSetup
Quick Start
git clonedocker compose uplocalhost:3000
AI Navigation

Your AI assistant knows your business

Just say what you need — ZenAI finds, combines, and delivers.

It's Monday afternoon. The quarterly presentation must be ready by tomorrow. Before: 3 hours of gathering data, building slides, checking numbers. Today: One sentence is enough.

You say:

Create the quarterly presentation for the board with all revenue figures, open projects, and team updates.

ZenAI automatically searches:
  • Q1 revenue report from the CRM
  • 47 active projects
  • 12 team updates from last week
  • Pipeline forecast
ZenAI asks:

Should I include the open proposals (€340,000) in the forecast?

Yes, include everythingQ1 closings only
3 minutes later:

Your quarterly presentation is ready

  • 14 slides in corporate design
  • All figures current and linked
  • Executive summary on slide 1
  • PDF + PPTX for download

Time comparison: Without vs. With ZenAI

Searching data45 min.0 min.Compiling60 min.2 min.Formatting30 min.1 min.Proofreading20 min.0 min.Total155 min.3 min.

Create a proposal for Mueller GmbH based on our last conversation and the current price list.

Finished proposal in 4 minutes — personalized, with correct prices.

4 min. instead of 45 min.

Your Data. Your Property.

At Your Location

ZenAI runs on your own hardware. No cloud dependency.

No Cloud Risk

Your data never leaves your network. GDPR-compliant by design.

You Decide

Full control over AI models, data sources, and access rights.

EU AI Act ReadyGDPR CompliantSelf-HostedOpen Source CoreISO 27001 Ready
Try ZenAI for FreeWatch Live Demo

No credit card · No IT knowledge required · Active in 12 minutes

The Problem

AI tools that forget. Every day.

You explain the same things to your AI over and over. Your context is lost after every session. Your knowledge is spread across 5 different tools.

No Memory

ChatGPT, Claude, Gemini — every session starts from zero. You are the context manager for your own AI.

"What was my project setup again?" — For the 47th time.

Vendor Lock-in

Your data lives on US servers. No export. No self-hosting. A compliance nightmare for every EU company.

"Where exactly is our data?" — No answer.

Tool Chaos

Notion for knowledge, ChatGPT for AI, Obsidian for notes, Pipedrive for CRM. None of it talks to each other.

5 tabs, 5 contexts, 0 connections.
How It Works

From Question to Insight

1
💬
Ask a Question
In natural language — text, voice, or image
2
🧠
AI Thinks
6 cognitive pillars analyze in concert
3
🔍
Retrieve Knowledge
From your memory, documents, and the web
4
Answer & Learn
ZenAI responds — and remembers the context
1
💬

Ask a Question

In natural language — text, voice, or image

2
🧠

AI Thinks

6 cognitive pillars analyze in concert

3
🔍

Retrieve Knowledge

From your memory, documents, and the web

4

Answer & Learn

ZenAI responds — and remembers the context

The Solution

One System. Six Pillars. Your Infrastructure.

ZenAI combines AI chat, knowledge management, agents, and long-term memory in one system — that you host yourself.

7-Layer Memoryfrom working memory to long-term storage
55+ AI Toolsin a single chat interface
Self-hostedDocker Compose, one command, your hardware
Open Source CoreZenBrain is Apache 2.0
ZenAI Architecture
Web UI · CLI · Mobile
React + Terminal + PWA
6 Cognitive Pillars
Memory · Reasoning · Agents · Output · Curiosity · Meta
ZenBrain OSS
7-Layer Memory · FSRS · Hebbian KG · Sleep Consolidation
PostgreSQL + pgvector
4 Contexts · HNSW · Self-hosted
Cognitive Architecture

Six Systems. One Brain.

Each pillar solves a concrete problem. Together they form the first AI with true cognitive depth.

Memory

Remembers everything — and knows what matters.

7-Layer ArchitectureFSRS RepetitionHebbian KGSleep Consolidation

Reasoning

Thinks in connections — not just tokens.

Global WorkspaceMulti-Hop InferenceFact CheckerChain-of-Thought

Agents

Acts autonomously — and stops when you want.

Persistent LoopsTool CompositionDynamic TeamsDebate Protocol

Output

Delivers documents — not just text in chat.

PPTXXLSXPDFDOCXCLI Agent

Curiosity

Asks on its own — instead of just answering.

Gap DetectionHypothesis EngineInfo Gain Scoring

Meta-Cognition

Knows its limits — honest instead of hallucinating.

CalibrationCapability ModelFeedback Loop
Comparison

Why not just use ChatGPT?

Because no other system meets all four requirements at the same time.

At a Glance

MemoryPrivacyToolsSelf-HostingOpen SourceAgents
ZenAI
ChatGPT
Notion AI
ZenAIChatGPTNotion AIMem AIOpen WebUI
7-Layer Memory~
Self-Hosted
GDPR-native~
55+ AI Tools~~
Persistent Agents~
Knowledge Graph~
Docs Export (PPTX/PDF)~
Open Source Core
Price from€0 / Free€20/mo€20/user$12/mo€0 / OSS
Use Cases

The right setup for every workflow.

From solo developer to enterprise team — ZenAI adapts.

Solo Developer
Full-Stack, personal projects

ZenAI remembers my entire project setup. I say 'like last time' and it knows exactly what I mean.

MemoryCLI AgentCode ExecutionGitHub Tools
Knowledge Worker
Research, Content, Analysis

My knowledge graph grows with every conversation. After 3 months, ZenAI has a better picture of my work than I do.

Knowledge GraphRAG PipelineDocs ExportWeb Search
Team Lead
5-15 people, SaaS

Agent teams handle our weekly reports. I just review — instead of writing.

Agent TeamsShared MemoryRBACGovernance
Enterprise IT
Compliance-required, DACH

Self-hosted, audit logs, EU AI Act ready. The only AI system our CISO signed off on.

Self-HostedAudit LogsField EncryptionEU AI Act
V4 Technology

7 Technologies That Make the Difference

Under the hood, technologies you won't find anywhere else — explained in your language.

AG-UI

AI That Shows Its Thinking

See in real-time what ZenAI is doing: which data it searches, which connections it finds, how it reaches an answer. No black box.

100% Transparency
Multi-LLM Routing

The Right AI Brain for Every Task

A fast model for simple questions, a powerful one for complex analyses. ZenAI chooses automatically — you never pay for more than needed.

3 AI Models
GraphRAG

Memory That Understands Connections

Your projects, clients, and documents aren't isolated files — ZenAI understands the relationships between everything. Like an employee who's been there for 10 years.

3-Layer Architecture
HyperAgents

AI That Improves Daily

ZenAI analyzes what worked and what didn't. Every day it gets a little better at helping you. Without you doing anything.

Self-optimizing
Adaptive Thinking

Deep Thinking — Only When Needed

For 'What's the weather?' it doesn't need 30 seconds of analysis. For 'Create a business plan' it does. ZenAI adjusts its thinking depth automatically.

Smart Budgeting
FSRS

AI That Never Forgets

Important information stays important. ZenAI uses the same algorithm that the best learning apps use — adapted for enterprise knowledge.

Neuroscience-based
Integration

Fully Integrated Into Your IT

REST API, MCP Server, CLI, A2A Protocol — ZenAI adapts to your existing infrastructure, not the other way around. And because the core is open source, you can inspect everything.

55+ Tools

100% Self-Hosted

Your data stays with you. No cloud dependency. Full control.

Open Source Core

ZenBrain — 12 neuroscience algorithms. Apache 2.0 licensed.

Continuous Improvement

ZenAI gets better every day. Automatically. Without installing updates.

Pricing

Start for free. Scale when you're ready.

No credit card lock. 14-day Pro trial included.

Free
€0
forever

For getting started. One context, core features.

  • 1 Context
  • 500 Memories
  • 50 Chats/month
  • Core AI Tools
Start for Free
Team
€79
user / month

Multi-user, shared memory, governance.

  • Everything in Pro
  • Multi-User + RBAC
  • Shared Memory
  • API Access
  • Governance Dashboard
Contact us
Enterprise
From €499
month flat

Self-hosted. SLA. EU AI Act compliance.

  • Everything in Team
  • Self-hosted Setup
  • Custom AI Models
  • Audit Logs
  • SLA + Support
Request demo

Annual plan: 2 months free. All prices excl. VAT.

Data Sovereignty

Your Data. Your Rules.

No US cloud risk. No vendor lock-in. No compromise.

Self-Hosted First

Your data never leaves your infrastructure. Docker Compose on your server.

GDPR-native

No workaround, no DPA with US providers. Privacy by design.

EU AI Act ready

Audit logs, governance dashboard, compliance tools. Ready for regulation.

Open Source Core

ZenBrain is Apache 2.0. Inspect every algorithm. No black-box AI.

Frequently Asked Questions

What you want to know before you start.

Do I need my own servers?

No. The Free and Pro tiers run on our infrastructure. Self-hosting is optional and available for Team/Enterprise. All you need: Docker and an Anthropic API key.

What happens to my data?

With self-hosting: nothing leaves your server. With our cloud: EU data centers, GDPR-compliant, no training on your data. You can export or delete all data at any time.

Which AI model does ZenAI use?

Primarily Claude (Anthropic). Fallback to Mistral (cloud) or Ollama (local). You can switch models in settings — including fully local inference without internet.

What is the difference between ZenAI and ZenBrain OSS?

ZenBrain is the extracted cognitive core — the memory algorithms as an npm package. ZenAI is the complete operating system that uses ZenBrain. Developers take ZenBrain, everyone else takes ZenAI.

Is there an API?

Yes. REST API + MCP Server. Full API access from the Team tier. The MCP Server enables integration with VS Code, Cursor, and other IDEs.

Can I migrate from ChatGPT/Notion?

We are working on import tools. Currently: Markdown import for notes and ideas. ChatGPT history import is on the roadmap for Q3 2026.

Ready for AI that
remembers you?

Start for free. Self-host when you're ready. No vendor lock-in — just your knowledge, permanently.

Start for FreeGitHub
git clonedocker compose uplocalhost:3000