Both Authentik and Keycloak provide enterprise-grade single sign-on, OIDC, SAML, and MFA for self-hosted application stacks. Both are serious identity providers — not lightweight utilities. The difference is philosophy: Authentik is built for modern DevOps teams who want something that works out of the box; Keycloak is the battle-tested choice for enterprise environments that need maximum protocol coverage and production history going back a decade.
Auth0 Essentials starts at $23/month for up to 1,000 monthly active users. Beyond that, the pricing scales by MAU in a way that surprises teams that grow: 10,000 MAUs cost roughly $240/month, and adding enterprise features (SAML, custom domains, MFA) pushes you to higher tiers. A self-hosted identity provider on a Liquid Web 4–8 GB Managed VPS eliminates per-MAU pricing entirely.
One docker-compose.yml deploys Authentik (SSO) + Outline (wiki) + shared PostgreSQL + Redis + Caddy; Coolify installs separately via its own script
Measured idle RAM: ~2 GB (Authentik + Outline stack) + ~770 MB (Coolify stack) = ~2.8 GB total across the host
Minimum Liquid Web tier: 8 GB Managed VPS (~$33–$40/mo); 16 GB recommended if you plan to deploy many apps via Coolify
Replaces: Okta (~$20/mo for 10 users) + Vercel Pro ($20/mo) + Notion Team ($160/mo for 10 users) = $200/mo down to ~$33/mo
Most engineering teams pay three separate vendors with no shared identity layer tying them together: Okta handles SSO, Vercel handles deployments, and Notion holds documentation. Each has its own login, its own access controls, and its own billing. This guide deploys the open-source equivalents — Authentik for identity, Coolify for self-hosted PaaS deployments, and Outline for team wiki — on a single Liquid Web 8 GB VPS, connected through OIDC so all three share one set of credentials.
One docker-compose.yml deploys Authentik (SSO) + Plausible (analytics) + Postal (SMTP) + shared PostgreSQL + MySQL + ClickHouse + RabbitMQ + Redis + Caddy; Coolify installs separately via its own script
Measured idle RAM: ~4.6 GB total across the host (our Compose stack + Coolify's stack)
Minimum Liquid Web tier: 16 GB Managed VPS (~$50/mo) — the 16 GB gives comfortable room for apps you deploy through Coolify
Vercel Pro ($20/mo) + Auth0 Essentials ($23/mo) + Postmark ($15/mo) = $58/mo just for foundations; this stack: ~$50/mo with unlimited apps, SSO, pageviews, and email
This is the "meta stack" — the infrastructure layer you deploy before anything else. Every other self-hosted app you run needs a place to live (Coolify), a way to log users in (Authentik), a way to measure traffic (Plausible), and a way to send email (Postal, the open-source SMTP platform). This guide wires all four together on a single Liquid Web 16 GB VPS as a reusable foundation.
Measured idle RAM: ~2.4 GB (full stack with Authentik); ~1.6 GB without Authentik
Minimum Liquid Web tier: 8 GB Managed VPS (~$33–$40/mo) — plus whatever storage you need for photos
Google One 2 TB: $9.99/mo. iCloud+ 2 TB: $9.99/mo. This stack: ~$33/mo VPS + full data ownership + face recognition that actually works offline
Google Photos removed its unlimited free tier in 2021, and iCloud keeps your photos locked inside Apple's ecosystem. Immich is a self-hosted photo and video library with a mobile app, automatic backup, face recognition, and a UI that genuinely resembles Google Photos. This guide deploys Immich on a Liquid Web 8 GB Managed VPS, with optional Authentik SSO for teams or families that want centralized login.
Minimum Liquid Web tier: 8 GB Managed VPS (~$33–$40/mo)
Linear (10 users) + Notion Team (10) + Okta (10): ~$260/mo. This stack: ~$33/mo
Engineering teams pay for Linear to track issues, Notion to write docs, and Okta to control who logs in to both. This guide deploys the self-hosted equivalents — Plane for project management, Outline for the team wiki, and Authentik as the OIDC identity provider — on a single Liquid Web 8 GB Managed VPS. One login covers everything.
One docker-compose.yml: Ollama + LiteLLM + Open WebUI + Authentik + PostgreSQL + Redis + Caddy
Team logs in via SSO (Authentik OIDC); per-team budgets enforced by LiteLLM
Runs Llama 3 70B at 40–60 tokens/sec on a Liquid Web L40S GPU server (48 GB VRAM)
ChatGPT Team (10 users): $250/mo — this stack on L40S: ~$1,040/mo 24/7 or ~$360/mo at 8 hr/day · 5 days/wk
Every prompt your team sends to ChatGPT Team or Copilot for Microsoft 365 leaves your network and is processed on OpenAI's or Microsoft's servers. For most teams that's an acceptable trade-off. For teams handling legal documents, source code, financial data, or customer PII, it is not. This guide deploys a private ChatGPT equivalent — Llama 3 70B running on your own GPU hardware, with a ChatGPT-quality web interface, your company's SSO, and per-team usage budgets — entirely inside your infrastructure.
