Nextcloud Hub: Files, Talk, Groupware, Office, Assistant, Flow
One self-hosted platform rather than a file-sync bolt-on: Files for sync/share and external storage; Talk for on-prem chat/video (optional SIP); Groupware for calendar/contacts/mail; Office for browser co-editing; Assistant for local AI tasks; Flow for automation. Limit: Talk scale and Office concurrency need extra backends/licenses for large orgs—budget High Performance Backend and editor capacity explicitly.
Self-hosted private cloud with clients and WebDAV access
Run the server on Linux with MySQL/MariaDB/PostgreSQL; users access via web UI plus desktop (Windows/macOS/Linux) and mobile (Android/iOS) clients. External storage connectors cover NFS, SMB/Windows Network Drive, SharePoint, S3/SWIFT, FTP and more so existing file systems stay under IT policy. Limit: you own patching, backup, HA, and capacity planning unless a partner hosts the instance.
Enterprise identity, sharing controls, and audit trails
LDAP/AD, native SAML 2.0, OpenID Connect, Kerberos, enforced MFA (TOTP, WebAuthn and others), password policies, file access control rules (IP, group, type, time), passworded/expiring shares, File Drop, video verification, remote wipe, and compliance-oriented activity logs. Benefits regulated teams that must prove who accessed what without sending files to a third-party SaaS tenant.
Encryption layers: TLS, server-side, optional E2EE
TLS for transport; optional AES-oriented server-side encryption for data at rest (including object storage scenarios); optional per-folder client-side end-to-end encryption with a published design whitepaper and enterprise options such as recovery keys/HSM identity issuance. Limit: E2EE is not a magic default for all Hub features—evaluate which apps and workflows remain usable when folders are E2EE-encrypted.
AGPLv3 open source with Enterprise support subscriptions
Server source is AGPLv3 on GitHub; Nextcloud states it does not ship proprietary open-core product modules. Community use is free; Enterprise plans (Standard/Premium/Ultimate) are seat-based subscriptions from a stated minimum user tier, buying support SLAs, longer maintenance, early patches, Guard, Global Scale options, and commercial connectors. Choose Enterprise when uptime and vendor SLAs matter more than pure DIY ops.
Local AI Assistant without mandatory cloud LLM tenancy
Assistant integrates summarization, translation, context chat over your data, and generation features inside Hub, designed to run with self-hosted or partner AI backends rather than forcing content into a public consumer AI product. Limit: model quality, GPU/CPU cost, and AI Act classification depend on how you deploy the models—treat AI as an optional module with its own DPIA.