No subscription. Your own server.

No. The SpaceBuilder price is for the software license. Install it on a PC you already have, an office server or a cloud server you rent, and files are stored on that server's disk. Server rental, domain and network costs are separate.

For small-scale use centered on files and chat, 2GB of RAM or more is enough to start. No GPU is needed, and for the CPU, check that the installer and operating system match before you look at the generation. Memory for the operating system and disk space for your files come on top. Leave headroom as concurrent users or document and image conversion grow. Running local AI on the same PC takes extra memory for the models. LM Studio recommends 16GB of RAM or more, and x64 CPUs need AVX2 support.

• LM Studio CPU and memory requirements

Free is for up to 2 people with no time limit, and the free license renews every month. It includes file management, document sharing, chat, a basic calendar and connections to existing mail accounts. Admin features such as changing the organization name and logo and audit logs start with Team.

A lifetime license is the right to keep using the version you bought without a monthly fee. You update by installing new versions yourself. Server costs, external AI charges and the SetFN business mail subscription are separate.

If up to 100 people share one server, choose Team. Club has the same admin features as Team with no limit on users registered on one server, and it runs 2 instances on one device, so you can split production and development. To run several servers under one license, look at Enterprise, which allows up to 5 and adds security and session management plus installation support. If you need more servers, custom development or data migration, contact us about Enterprise+.

Concurrent devices are the server devices one license runs on at the same time: 1 for Free, Team and Club, up to 5 for Enterprise, and by agreement for Enterprise+. An instance is a SpaceBuilder server running on one device: 1 for Free and Team, 2 for Club, and up to 10 per device for Enterprise. If you change PCs or the device ID changes, choose "Release license" on the old server first, then register the new environment. You can change devices 5 times by default, and the remaining count is shown under "License management" in your account. When you reach the limit, send a request to adjust the device change limit.

The SetFN Space app is free. Add several workspace addresses to one app, personal, company, club and so on, and sign in to each with the account you have there. There is no limit on how many workspaces you can add.

A plan's user count is the number of accounts you can register. Concurrent connection tests passed at 10,000 connections for workspace file management and 5,000 for Webspace. What Webspace actually handles depends on database performance and server specs.

Look at how much you will store and how fast it needs to be read and written. If file browsing, uploads and database work often overlap, an SSD is the better choice. If you move large files often, check the server's network bandwidth as well as disk speed.

Files, conversations and member information are stored on the workspace server you run. The operator decides where they are stored and who can access them. If you connect an external AI or mail service, the data that feature needs may go out to that service.

Whoever runs the workspace server handles backup and recovery. Back up the original files, database and settings together, and keep the copy on storage separate from the server. Test once that you can actually restore from it.

Yes. Prepare a server address reachable from outside and an HTTPS connection, and you can connect your own domain too. For office-only use, limit access to the internal network or a VPN. If a PC is your server, it has to stay powered on and online while people connect from outside.

• External access, domain and HTTPS setup

Workspace AI uses local models by default. Run Ollama or LM Studio on the server or a connected PC, then register the runtime address and model in the workspace. If you have no GPU or cannot run a large model, sign in to Ollama and use an Ollama Cloud model. Ollama paid subscriptions and usage credits are separate from SetFN pricing, and other external AI APIs charge their own fees.

• Ollama Cloud guide
• Ollama subscription and usage pricing

If you plan to run local AI on a GPU, we recommend 16GB of VRAM or more as a starting point. Small models run with less VRAM or on the CPU alone, and the memory needed depends on model size and quantization. For both NVIDIA and AMD, check Ollama's supported GPU list and the operating system and driver requirements. With 4-bit quantized models, an RTX 3090 (24GB) is a workable setup for Gemma 4 26B and an RTX 5090 (32GB) for models around 30B. Long documents, concurrent requests and an embedding model running alongside need more memory headroom.

• Ollama supported GPUs and drivers
• Gemma model memory guide

Besides the AI model that writes answers, you need an embedding model. It turns document content into vectors, the numeric data used for search. Load an embedding model such as bge-m3 in a local runtime like Ollama, then pick the model and vector store in the knowledge indexing settings and index your documents. Once indexing is done, the AI finds related documents and uses them as the basis for its answers. That is RAG. Even if you use Ollama Cloud for answers, the embedding model and document index are set up separately.

• SetFN knowledge indexing setup
• Ollama embedding models

IMAP and POP3 connections to existing accounts such as Gmail are free. SetFN business mail is a separate subscription: a broker server receives your mail and passes it to your workspace server over an mTLS (mutual authentication) connection. Mail your workspace server could not receive is held on the broker server for up to one week.

Team and Club get online technical support. Enterprise includes installation, server and domain connection and direct technical support. For data migration, custom development, managed operations or a service level agreement (SLA), contact us about Enterprise+. The scope of support and the contract terms are settled before purchase.