What is niil?
niil is a secure, EU-sovereign AI workspace. It brings three things together in one place:
- Secure Chat — a fast, multi-model chat for everyday work, grounded on your own files and knowledge bases when you need it.
- Agents — reusable assistants you configure once (model, instructions, tools, knowledge) and then use in chat or in workflows.
- Workflows — a visual canvas for connecting agents, tools and steps into repeatable, multi-agent processes.
What makes niil different is where and how your data is handled:
- EU-first. Your organization uses only EU-processed (or self-hosted) models by default. See EU sovereignty & residency.
- Encrypted. Messages, documents and credentials are encrypted at rest; each conversation has its own key. See Encryption.
- Self-contained. No third-party analytics or telemetry; the AI tools run in niil's own runtime. See No external vendors.
Who uses which parts
| Role | Can do |
|---|---|
| Everyone | Secure Chat, create & use agents, build & run workflows, knowledge bases, see their own usage, manage their profile. |
| Organization admins | Everything above, plus invite/remove members, set spending limits, configure connections, manage plans & billing, and set org security policy (EU-only models, login MFA). |
System operators (the people who run the niil platform itself) have a separate operator console that is not covered in this guide.
Where to go next
- New here? Start with Sign up & sign in, then take the Interface tour.
- Want to chat right away? See Chatting basics.
- Building automation? Jump to Agents and Workflows.
- Evaluating security? Read the Trust overview.