Next-generation agent platformthat actually gets work done.

huoku is the sovereign agentic platform that aggregates your organizational knowledge and puts it to work. AI agents that don't just answer questions: they send emails, run workflows, call APIs, and automate real business processes. On a dedicated cluster on European infrastructure, or fully on-premise.

Deployment
EU cloud · On-premise
Components
Agents · Workflows
Security
Document-level
Problem

Your knowledge is trapped. Your processes are waiting.

Knowledge exists but can't be turned into action. AI automation gets blocked by security teams. Companies that rush into shared AI platforms trade control for convenience, and can't get it back.

01

Scattered knowledge

Company knowledge lives in Confluence, SharePoint, network drives, email, and dozens of other systems. Employees waste hours searching. Worse, none of that knowledge is available to any automated process.

02

The security trap

Multi-tenant AI platforms pool your data into shared infrastructure, often under foreign jurisdiction. For regulated industries and security-conscious companies, that's a non-starter. Your intellectual property, customer data, and compliance-sensitive information belong in an environment dedicated to you and governed by European law.

03

Dangerous dependencies

When business-critical processes depend on services you can't control, their outage becomes your outage. You need infrastructure that's yours, where every dependency is chosen, not imposed.

Solution

From scattered knowledge to automated action.

huoku closes the gap between knowing and doing: satellite crawlers continuously aggregate knowledge from all your systems into a unified intelligence layer, and configurable AI agents use that knowledge to perform real work, triggered by events, schedules, or people.

01 · Aggregate

Knowledge from every system

Satellite crawlers continuously gather knowledge from Confluence, SharePoint, file shares, and other enterprise systems into a centralized vector database.

02 · Reason

Context-aware decisions

AI agents access all organizational knowledge, understand context, and make informed decisions about what needs to happen.

03 · Act

Real work, not just answers

Agents deliver results through connectors: they send emails, call APIs, publish events, and execute multi-step workflows.

How it works

Trigger. Agent. Action.

Every automation in huoku follows the same pattern. A trigger starts the workflow. An agent reasons with your company's knowledge. An action delivers the result.

01 · Trigger

Anything can start a workflow

Incoming emails, Slack messages and mentions, REST API calls, Kafka events, webhooks, schedules, changes on monitored web pages, or new documents arriving in the index. Results flow back where the trigger came from: an agent can reply directly in the Slack thread that mentioned it.

EmailSlackRESTKafkaWebhookScheduleWeb pagesNew documents
02 · Agent

Reasons with your knowledge

Each agent is independently configured with its own knowledge access, skills (multi-step workflows), tools (actions via connectors), and even sub-agents for complex tasks. Admins create and manage agents through a visual interface, no coding required.

KnowledgeSkillsToolsSub-agents
03 · Action

Delivers the result

Agents act through the same connectors that trigger them. Send an email response, call a REST endpoint, publish to Kafka, fire a webhook.

EmailRESTKafkaWebhook
Use cases

Agents in production.

Sensor data and manuals, together. Anomaly detected, the right team alerted, work order opened. Same platform, on a cluster dedicated to you.

Industrial · Running 24/7

Predictive maintenance.

A scheduled workflow continuously analyzes machine state. An anomaly is detected before anyone has time to ask, and the corrective action starts in the background.

01Aggregate

Sensor data + manuals

The agent fuses OPC-UA streams, MES history, and PDF manuals into one context.

DATA.IN / 4 SOURCES
  • OPC-UA sensor streams
  • MES production history
  • PDF maintenance manuals
OPC-UA · 12.4k/s
02Reason

Model diagnosis

Spots the anomaly, retrieves prior cases, and proposes root causes with source citations.

ANOMALY · LINE 03
  • vibration +38%
  • temperature ↗ 87 °C
  • ≈ manual §4.2
confidence 0.91
03Alert

Route to the right team

Routing follows identity and shift: Teams, email, or a work order, directly.

TEAMS · #MAINTENANCE
  • @shift-tech
  • Line 03 — predictive
  • open source →
response 3 min
04Act

Work order + instructions

Creates a maintenance order in the ERP, attaches step-by-step instructions, closes the loop.

SAP · WO-21847
  • ✓ order opened
  • ✓ instructions attached
  • ✓ window 14:00
status active
Procurement · Contracts

Procurement assistant.

The agent tracks procurement announcements nightly and compares contracts and offers side by side: price, coverage, terms, and special clauses in a single view, with citations to the source documents.

  • Side-by-side comparison table with rows extracted automatically
  • Citations to the original PDFs always visible
  • Your team reviews a ready analysis instead of digging through PDFs
Knowledge · Chat

Internal knowledge base.

Employees ask an agent in chat, and it finds the answer from anywhere in the organization: Confluence, SharePoint, network drives, intranet pages. Answers come with source citations, and everyone sees only what they could open in the source system.

HR · Daily

HR reporting automation.

A daily agent aggregates sick-leave notifications from incoming messages, compiles a summary, and delivers it to HR and team leads. A recurring manual chore becomes a report that is simply there every morning.

Product

The control room.

Agents, workflows, identity, monitoring, and tracing inside the same admin panel. One window for operators, security, and audit.

Admin panel · AI agents

Manage AI agents centrally.

One view for all your AI agents. Each card shows the agent's role, available tools, and live status.

  • 01Agent overviewCard view shows the agent's role, tool count, and status at a glance.
  • 02Interface tagsColor-coded tags (chat, rest, kafka) immediately indicate which interfaces an agent supports.
  • 03New agent creationAdd agents in a few clicks, with no programming and no DevOps tickets.
Manage AI agents centrally.
Document-level security

Permissions follow the data, from chunk to answer.

Security is part of the architecture: every chunk, query, and answer carries an identity.

01 · Access control

Chunk metadata

Every chunk preserves the original ACLs: groups, users, and visibility tags.

02 · Sanitization

Data Guard

Regex-based PII masking before storage. Personal data never enters the index.

03 · Search · RAG

Vector database

Hybrid search returns only the chunks the asker is authorized to see. The answer cannot leak data.

04 · Audit

Tracing

Every query, retrieval, and answer is traceable down to the user and the source.

PII redaction before AI sees it

Sensitive identifiers are automatically masked in the document pipeline, so GDPR and EU AI Act are satisfied without giving up the AI value.

GDPREU AI ActData Guard
Source

Patient Matti Meikäläinen, SSN 120464-123A, arrived at the clinic on 2026-04-12.

For AI

Patient Matti Meikäläinen, SSN [REDACTED_FI_SSN], arrived at the clinic on 2026-04-12.

The result: the AI only answers from data the user is permitted to see, without unbundling or duplicating your existing access model.

Key capabilities

Built for enterprise. Controlled by you.

Everything an enterprise needs to turn knowledge into automated action, configurable from one admin panel.

Knowledge aggregation
Satellite crawlers continuously index Confluence, Jira, SharePoint, Teams, Slack, SQL databases, websites, file shares, and more. 10+ document formats including OCR.
Configurable agents
Admins create agents through a visual interface. Custom prompts, knowledge access, skills, tools, triggers, all configurable per agent.
Multiple triggers
Email, Slack, REST, Kafka, webhooks, schedules, monitored web pages, new documents. Flexible automation entry points.
Hybrid search
Keyword, semantic, and hybrid search with AI reranking. Source attribution on every result.
Multi-agent system
Agents work as teams: sequential assembly lines, free-form groups, or coordinated by a supervisor agent.
Human oversight
Approval gates for sensitive actions, live view of running agents, stop any run instantly, full audit trail.
EU AI Act ready
Prohibited uses (Article 5) blocked automatically. Auto-generated system cards per agent. Configurable audit retention.
Multiple interfaces
Chat UI for employees. REST API and Kafka for integrations. MCP and A2A protocols for AI interoperability.
Central admin panel
One panel for agents, identities, knowledge collections, security, monitoring, and traces.
Model agnostic
Your model, your choice: EU-hosted open models by default, fully local on your own hardware, or any external provider. Switch at runtime.
Deployment choice
Always a dedicated single-tenant cluster: on European cloud, in your own cloud account, or fully on-premise.
Enterprise security
Keycloak SSO, role-based access, document-level permissions, Data Guard PII masking, encryption everywhere, full audit trail.
Deployment

Sovereign by design. Deployed your way.

huoku runs as a Kubernetes application on a cluster dedicated to you. Never shared, never multi-tenant. Most customers run on European cloud infrastructure with EU-hosted models: data stays in Europe, under European jurisdiction, in a cluster that's yours alone. And when your requirements demand it, the entire platform installs on-premise with fully local models, a completely air-gapped AI platform with zero external communication.

Default

European cloud

Dedicated cluster on European infrastructure, operated by us or running in your own cloud account.

Your cloud

Your infrastructure

Deploy into your existing Kubernetes, in any cloud or data center.

Air-gapped

On-premise

Local models via Ollama, zero external communication, complete isolation.

Key points
  • Always single-tenant: one customer, one cluster
  • No mandatory dependencies outside your cluster: the only external call is to your chosen model provider, and even that can move inside with local models
  • Single Helm chart deployment
  • No GPU required: embeddings run on CPU
  • Horizontal scaling for growing workloads
  • Automated TLS certificate management
  • Health monitoring and observability built in
Track record

Built by people who've done this before.

huoku's roots go back to 2015: the founding team delivered real-time data analysis and AI solutions for international telecom operators, IoT customers, and Finnish public sector organizations. huoku is the product of that experience, purpose-built for companies that need their AI to work reliably, securely, and independently.

European AI, backed by Europe

huoku has been selected for funding through the European Union's AI-on-Demand (AIoD) programme.

If it doesn't run reliably on the customer's own infrastructure, it doesn't run at all.
Ville Seppälä·CTO · huoku
Start with a conversation

AI on a cluster that's yours alone.
On your terms.

Tell us about your workflows. Together we'll identify 2-3 processes worth automating and deploy a pilot on a dedicated cluster, on European cloud or your own infrastructure. Agents work on real tasks within weeks. We reply within two business days.

Single-tenant · European infrastructure · Production-grade