100% in Finland
Your data stays in Finland and within your organisation.
ISO 27001
Information security practices in line with the certification.
No training on your data
Your data is not used to train AI models.
Ekokompassi-certified
Measured carbon footprint and offset emissions.
We bring together AI models, development tools, management and high-performance GPU infrastructure into one managed solution. Manage resources and usage securely in a private Finnish AI environment.
We bring together AI models, development tools, management and high-performance GPU infrastructure into one managed solution. Manage resources and usage securely in a private Finnish AI environment.

Models and pricing
We select, validate and maintain the platform’s model library based on customer needs.
Pricing consists of a monthly fee of €200/month per customer environment, plus usage-based token costs of the AI models.
Model |
Status
|
Input token
|
Output token
|
|
GLM-5.3-Flash |
Active |
0.25€/Mtok |
0.50€/Mtok |
|
Qwen3.8-27B-FP8 |
Active |
0.45€/MTok |
3.00€/MTok |
|
dots.ocr |
Active |
0.05€/Mtok |
0.10€/Mtok |
| Different models of the Mistral |
|
|
|
|
Different models of the Kimi K
|
|
|
|
Pricing
Pricing consists of a monthly fee specific to the customer’s environment, as well as token costs determined by the use of language models. You can monitor token usage in real time from the platform’s admin dashboard, which also makes cost management transparent.
Model |
Status |
Input token |
Output token |
|
GLM-5.3-Flash |
Active |
0.25€/Mtok |
0.50€/Mtok |
|
Qwen3.8-27B-FP8 |
Active |
0.45€/MTok |
3.00€/MTok |
| Different models of the Mistral |
|
|
|
|
Kimi K 2.5
|
|
|
|
For administrators and management
Management that keeps you in control.
Users and access rights
User-, group- and resource-level permissions for models, API keys, RAGs, MCP and Preset Agent tools.
Usage monitoring
Real-time token consumption by key and by model, so costs can be allocated to cost centres or teams, for example.
Managed platform
A fully managed service that lets you use AI resources without worrying about the underlying infrastructure.
For developers
APIs for different use cases.
Kaikki mitä tarvitset AI-mallien käyttöön, agentteihin tai oman datan hyödyntämiseen samassa ympäristössä ja samalla API-avaimella.
AI models
An up-to-date selection of AI models for a variety of use cases, fully maintained. Other models and models trained for a specific use case can also be added to the AI platform.
API and application integrations
Resources for AI agents
Responses API endpoint for building agentic AI solutions and calling tools and models. Built-in MCP connects agents to external tools and systems.
MCP
Built-in MCP connects agents to external tools and systems.
RAG
RAG tools for bringing your organisation's data onto the platform. Separate APIs for using RAG.
API Keys
Scoped API keys that can be linked to models, RAG, Preset agents, external models (proxy), users and user groups. Token usage tracking per API key.
Architecture
How your data flows, and where it never goes.
Every customer has their own isolated environment where data does not leave. All computing takes place in Magic Cloud’s data centre.
Use cases
What can be built on the platform?

Software Development
Use for code generation or integrate AI models into your software.

Document Intelligence
Search, summarize, and analyze documents using AI.

AI agents
Automate workflows and business processes.

Internal AI Assistant
Secure AI assistant trained on company knowledge.

Customer Service AI
Automate support with your own data.

Software Development
Use for code generation or integrate AI models into your software.

Document Analysis
Search, summarize, and analyze documents using AI.

AI agents
Automate workflows and business processes.

Internal AI Assistant
Secure AI assistant trained on company knowledge.

Customer Service AI
Automate support with your own data.
Case Study
Security and compliance
Security is essential.
Our solution is designed to meet EU and Finnish regulatory requirements and to satisfy even the most demanding security and compliance needs. We have compiled a description of the AI platform's security, data processing and compliance into a single document.
- Managed service and division of responsibilities
- Platform security
- Data management and data protection
- Compliance: ISO 27001 and NIS2
- Continuity and incident management
- Sustainability
Security and compliance
Security is essential.
Our solution is designed to meet EU and Finnish regulatory requirements and to satisfy even the most demanding security and compliance needs. We have compiled a description of the AI platform's security, data processing and compliance into a single document.
Download the Trust Pack (PDF)
Two ways to use the AI platform
AI development platform
Ready-made packageThe fastest route to working AI solutions
GPU capacity, AI models and tools (MCP, RAG, API) in one package
Model lifecycle management and full service maintenance included
Flexible scalability and a trial period without a long-term commitment
Especially suitable forOrganisations that want to deploy AI solutions quickly.
Dedicated GPU
Reserved capacityReserve capacity for demanding AI workloads and your own development tools
GPU capacity reserved for the customer and predictable performance
Stronger isolation and customer-specific security provisions
Your own tools and models – AI development tools and application layer available as an add-on service
Especially suitable forOrganisations that have their own AI solutions and development expertise.
Deployment
Get started in days, not months.
Deploying the AI platform is fast. Book a free demo meeting or start the deployment by filling in the form on the right.
Submitting the form does not commit you to an order. We will go through your needs with you and stay in touch about the next steps.
Let’s discuss your AI platform needs
In the free demo meeting, we will go through the features of the AI platform in the context of your use cases.
Frequently asked questions
A managed AI platform provides companies with a complete environment for leveraging generative AI. The service includes a customised isolated computing environment, pre-built generative language models, powerful GPU computing capacity, managed data management and integrations with existing systems. The platform enables rapid deployment, agent-based functionality and Retrieval Augmented Generation (RAG) functionality, as well as standard API interfaces for business needs.
Where is the data stored?
All data is stored in Magic Cloud’s own data centers located in Finland. The platform operates entirely on domestic infrastructure, so the data remains within Finland’s borders.
Can data be transferred out of Finland?
No. All data is processed in a secure environment at data centers located in Finland and is not transferred outside of Finland. Furthermore, the language models used do not utilize customer data to train their own models.
What language templates are available on the platform?
The available language models can be found in our current price list. In addition to the pre-built models, open-source language models—such as those available from Hugging Face—can be imported into the platform according to the customer’s needs.
Can I import my own models to the platform?
Yes. The platform also supports the customer’s own open-source language models, allowing the solution to be tailored to the organization’s needs.
How quickly can the platform be put into use?
The deployment schedule depends on the use case. At its simplest, the platform can be deployed with just a few clicks, while more extensive and customized solutions are typically completed within a few days.
How has security been taken into account in the AI platform?
-
The service is hosted in Magic Cloud's home cloud, ensuring data sovereignty and data retention.
-
The computing environments are isolated and protected.
-
The platform includes security and backup functions such as updates, vulnerability monitoring and control.
-
All data traffic takes place over secure connections, and customer data is never out of your control.
How are prices determined?
Pricing consists of a monthly fee specific to the customer’s environment, as well as token costs determined by the use of language models. You can monitor token usage in real time from the platform’s admin dashboard, which also makes cost management transparent.
How does the computing capacity scale?
GPU computing capacity scales according to your needs, ensuring performance even at high model sizes and workloads. The platform automatically optimises resource usage and the customer only pays for actual usage, making the service cost-effective.
Are you interested in our AI platform? Request a quote easily here!
Contact us, and we’ll design a solution tailored to your organization. You can request a quote or start the implementation right away if you already know what you need. If, on the other hand, you’d like to brainstorm options or have any questions, we’d be happy to discuss them with you.
Are you interested in our AI platform? Request a quote easily here!
Contact us, and we’ll design a solution tailored to your organization. You can request a quote or start the implementation right away if you already know what you need. If, on the other hand, you’d like to brainstorm options or have any questions, we’d be happy to discuss them with you.


