“De-Google” fails when people try to replace every Google product in one weekend. It succeeds when you migrate by job-to-be-done: files, photos, analytics, documents, then mail—accepting that some Google surfaces (Search, Maps, Android) may stay for years. Open source gives you control and predictable cost; it does not give you zero maintenance.
In 2026 the FOSS stack for personal and small-team Google replacements is mature enough to be boring—in a good way. Nextcloud and Seafile cover drive/sync; Immich and Ente cover photos; Plausible, Umami, and Matomo cover analytics; AppFlowy and AFFiNE cover docs/notes. Mail remains the hardest layer and should be optional until the rest is stable.
Start from pain, not ideology. If Drive sharing ACLs are the daily friction, fix files first. If photo library lock-in worries you, fix photos first. If GDPR and cookie banners dominate, fix analytics first. Parallel migrations multiply support load and hide which cutover actually worked.
Key takeaways
- Migrate one Google surface at a time; files or photos first for most households and teams.
- Nextcloud is the default Drive suite; Seafile wins when fast sync matters more than apps.
- Immich and Ente are the practical Photos shortlist—validate mobile backup early.
- Plausible/Umami for lightweight privacy analytics; Matomo when you need deeper ownership.
- AppFlowy and AFFiNE replace Docs/Notion-shaped work—not every Google Workspace feature.
- Self-host only what you can back up, update, and restore; use the stack builder to stay coherent.
Map Google jobs to open stacks
| Google job | Open shortlist | Difficulty | Notes |
|---|---|---|---|
| Drive / Docs files | Nextcloud, Seafile | Medium | Clients + sharing ACLs decide success more than web UI |
| Photos | Immich, Ente | Medium | Mobile backup and library size are the real tests |
| Analytics | Plausible, Umami, Matomo | Low–Medium | Easiest win for many sites; cookie posture improves immediately |
| Docs / notes | AppFlowy, AFFiNE | Medium | Realtime collab and import quality vary—pilot writers first |
| Mail (optional) | Stalwart / Mailcow / hosted FOSS | High | Deliverability and spam are ops-heavy; delay until ready |
| Search / Maps / Android | Partial / hybrid | High | Often stay hybrid; do not block the rest of the migration |
Files: Drive → Nextcloud or Seafile
Nextcloud
Open in catalogNextcloud is the broad suite: sync, share links, apps, and optional Collabora/OnlyOffice for browser editing. Strengths: you own the storage and ACLs; clients exist for major platforms. Limits: desktop sync edge cases still happen, office editing needs extra services, and you own backups. Choose Nextcloud for “mini Workspace”; choose Seafile when sync performance and simplicity beat an app ecosystem. See Google Drive alternatives.
- Best for
- Households and teams leaving Drive/Dropbox-style sync.
- Deployment
- Self-hosted (or trusted host).
- Pricing
- Open-source; storage and ops cost money.
- Unique
- Drive replacement starts with clients and sharing, not the web UI.
Seafile
Open in catalogSeafile focuses on fast, reliable file sync with a thinner collaboration surface than Nextcloud. Strengths: efficient sync and a clearer “just files” scope. Limits: fewer built-in collaboration apps; browser Office needs a separate strategy. Choose Seafile when sync performance and simplicity beat an app ecosystem.
- Best for
- Users who want fast sync without suite sprawl.
- Deployment
- Self-hosted (or trusted host).
- Pricing
- Open-source; storage and ops cost money.
- Unique
- Sync-first Drive alternative.
Photos: Google Photos → Immich or Ente
Immich
Open in catalogImmich is the popular self-hosted Photos-like stack with strong mobile backup momentum. Strengths: timeline UX, albums, and a credible path off Google’s library lock-in. Limits: face recognition and search quality differ from Google’s ML; large libraries need disk and patience; mobile battery/upload behavior needs a real pilot. Choose Immich for self-host control; choose Ente when encryption and managed simplicity weigh more. Details: Google Photos alternatives.
- Best for
- People with large personal libraries and phone auto-backup habits.
- Deployment
- Self-hosted.
- Pricing
- Open-source; hardware costs apply.
- Unique
- Mobile backup reliability is the acceptance test.
Ente
Open in catalogEnte emphasizes end-to-end encryption and a privacy-first cloud posture (including hosted options). Strengths: threat-model clarity and everyday mobile backup. Limits: self-host complexity differs from Immich’s mainstream homelab path. Choose Ente when encryption and managed simplicity weigh more.
- Best for
- Privacy-conscious users who may prefer managed encrypted hosting.
- Deployment
- Self-hosted or privacy-focused hosted.
- Pricing
- Open-source cores; hardware or subscription costs apply.
- Unique
- Encryption-first Photos alternative.
Analytics: GA → Plausible, Umami, or Matomo
Plausible
Open in catalogPlausible delivers lightweight, privacy-oriented web analytics without the GA cookie nightmare for many sites. Strengths: charts stakeholders read unaided. Limits: deliberately shallow versus Matomo. Read best Google Analytics alternatives.
- Best for
- Site owners reducing GA dependency with simple metrics.
- Deployment
- Cloud or self-hosted.
- Pricing
- Open-source; hosted tiers common.
- Unique
- High ROI, low-drama Google Analytics cut.
Umami
Open in catalogUmami delivers Docker-friendly, privacy-oriented pageviews with a modern multi-site dashboard. Strengths: easiest genuine self-host of the lightweight counters. Limits: no deep funnels—use Matomo when analytics is a regulated data store.
- Best for
- Teams that already run containers and want FOSS pageviews.
- Deployment
- Self-hosted.
- Pricing
- Open-source.
- Unique
- Low-ops privacy analytics.
Matomo
Open in catalogMatomo goes deeper when you need on-prem ownership, richer reports, or stricter data residency. Strengths: GA-like breadth under your infrastructure. Limits: more ops than Plausible/Umami.
- Best for
- Orgs needing deeper web analytics ownership.
- Deployment
- Self-hosted or cloud.
- Pricing
- Open-source core; hosted/premium options available.
- Unique
- Depth and control over minimal counters.
Docs: Docs/Keep → AppFlowy or AFFiNE
AppFlowy
Open in catalogAppFlowy covers notes, docs, and Notion-shaped productivity with local-first or self-host paths. Strengths: markdown-friendly workflows and escape from a single vendor wiki. Limits: realtime multiplayer polish trails Google Workspace; import from Docs is imperfect. Choose it for knowledge bases and personal productivity; keep Collabora/OnlyOffice beside Nextcloud if browser Office compatibility is mandatory.
- Best for
- Individuals and small teams replacing Docs/Notion-shaped work.
- Deployment
- Desktop / self-hosted / cloud variants.
- Pricing
- Open-source core; hosting varies.
- Unique
- Docs migration is a writing workflow change, not only file sync.
AFFiNE
Open in catalogAFFiNE covers docs, notes, and whiteboard-ish collaboration closer to Notion than to classic Google Docs. Strengths: flexible canvases and a FOSS path off vendor wikis. Limits: enterprise admin and import polish need a real pilot. Choose it when whiteboard+docs in one surface matters.
- Best for
- Teams wanting docs plus whiteboard-ish collaboration.
- Deployment
- Desktop / self-hosted / cloud variants.
- Pricing
- Open-source core; hosting varies.
- Unique
- Notion-shaped docs with canvas flexibility.
Mail: optional, last, and honest
Self-hosted mail is a deliverability and abuse problem disguised as a software install. Many successful de-Google stacks keep a reputable hosted mailbox (or a privacy-focused provider) while self-hosting files, photos, and analytics. If you insist on FOSS mail, budget for reverse DNS, SPF/DKIM/DMARC, spam filtering, and ongoing reputation monitoring—or accept that important mail will land in spam.
Migration playbook
- Inventory: list Google products you use weekly, not annually—migrate the weekly ones first.
- Week 1–2: analytics cutover on one site (fast confidence win).
- Week 3–6: Drive pilot—sync one shared folder tree; verify share links and offline clients.
- Month 2: Photos—enable mobile backup to Immich/Ente while keeping Google read-only.
- Month 3: Docs/notes for one team; freeze Google Docs for new work in that scope.
- Mail only after backups, monitoring, and a named owner exist for every prior layer.
Selection criteria
- Who owns backups and restore drills for each service?
- Mobile quality for the jobs that happen on phones (photos, chat later, drive offline).
- Sharing model: public links, passworded shares, SSO, guest accounts.
- Hardware and bandwidth honesty for photo/video libraries.
- Legal: data residency, retention, and export requirements.
- Exit plan: can you leave the FOSS tool later with your files intact?
Frequently asked questions
- Can I fully replace Google in 2026 with open source?
- You can replace Drive, Photos, Analytics, and much of Docs-shaped work with mature FOSS. Search, Maps, and Android usually stay hybrid. Treat “100% Google-free” as a long horizon, not a weekend project.
- What should I migrate first?
- Analytics for sites (quick win), then files or photos depending on which lock-in hurts more. Avoid starting with mail. Parallel big-bang cutovers are how de-Google projects stall.
- Nextcloud or Seafile for Drive?
- Nextcloud if you want a suite and apps; Seafile if fast sync and a thinner surface are enough. Pilot with your actual desktop clients and share-link habits before declaring a winner.
- Do I need to self-host everything?
- No. Many people self-host Immich and analytics while using managed Ente or a hosted Nextcloud provider. Own the threat model: self-host when control matters more than convenience.
- Where do I build a coherent stack?
- Use the stack builder and self-hosted catalog, then deep-dive Drive and Photos alternative pages for shortlists.
Conclusion
Replace Google by jobs, not by brand loyalty. Ship analytics, then files, then photos, then docs; treat mail as optional advanced mode. Pick boring, well-documented tools, name an owner for backups, and keep Google read-only until restore drills pass. Control compounds—each stable cutover makes the next one cheaper.
Build a stack for this use case.
Answer nine practical questions and compare three transparent architectures with costs, free limits, lock-in, and migration paths.
Build my stack
