This page is a verifiable use case. Everything we claim here is public and reproducible: every URL, every container, every command is open for you to check yourself.
The 30-second summary
d0a1.es runs on the same d0a1 stack it advertises. It’s not a staging example or a controlled demo: it’s the marketing site, the chat app, the Git repository, and the storage management platform — four subdomains in real production, powered by 30 Docker containers and 14 local AI models, paying zero euros in external tokens.
We audited the stack live on June 3, 2026 at 21:02 UTC. This page reflects that exact moment on the server.
The actual infrastructure
| Resource | Value |
|---|---|
| VPS | Clouding, 4 vCPU + 16 GB RAM flavor |
| OS | Ubuntu 24.04 LTS, kernel 6.8.0 |
| Disk | 99 GB total, 87 GB used (92 %) |
| Load average | 1.7 — stable, no saturation |
| Uptime | 5 days since last reboot |
| VPS cost | ~€25 / month |
What’s running inside (audited live)
30 Docker containers serving 4 subdomains, classified by function:
Local AI core — 8 containers
ollama— 14 models served onlocalhost:11434(9 local, 5 cloud fallback).ollama-copilot-proxy— HTTP proxy on:11435for editor integration.qdrant— vector database on:6333for semantic search.qdrant-mcp-server— MCP exposure on:8080/:8181/:9090.openwebui— chat interface atapp.d0a1.es.mcpo+mcpo-docker-socket-proxy— unified MCP gateway.wp-ai-fastapi— Python bridge between WordPress and Ollama on:8000.searxng— private metasearch on:8888.
WordPress + data — 4 containers
wordpress— d0a1.es (this site) on Apache + PHP 8.3.wordpress_db— MariaDB 11 with all pages, posts and CPTs.wpcli— WordPress CLI without UI, for operations.redis— object cache to reduce MariaDB queries.
Automation — 3 containers
n8n— visual workflows on:5678(RSS to post, social publishing, weekly reports).n8n_db— Postgres for n8n.redis(shared with WordPress).
Dev and Git — 4 containers
gitea— git.d0a1.es, 14 repos (6 public, 8 private).gitea-dind— Docker-in-Docker for Actions runners.gitea-act-runner— pipeline executor.workspace-git+workspace-artifacts— storage.
Mail and proxy — 5 containers
mailserver— Postfix + Dovecot + OpenDKIM. Sending and receivinghello@d0a1.esworking.snappymail— webmail on:8888.nginx(openresty) — reverse proxy on:80/:443for all 4 subdomains.terminal-proxyandterminal— web terminal on:7681.
Storeroom OS — 3 containers
storeroom-web— Next.js 16 + Tailwind v4 atstoreroom.d0a1.es:3002.storeroom-dbandstoreroom-os-db— dual Postgres (app and operations).storeroom-docuseal— digital signature for contracts.
Workspace agent — 2 containers
workspace-browser— Playwright headless for QA and screenshots.workspace-artifacts— agent execution storage.
AI models on the VPS
9 models served locally (17.5 GB on disk, €0 in tokens) plus 5 remote Ollama Cloud models as fallback for specific cases:
| Model | Size | Real use on d0a1.es |
|---|---|---|
| granite4.1:8b | 5.3 GB | General reasoning, draft review |
| qwen3.5:4b | 3.4 GB | SEO content generation |
| ministral-3:3b | 3.0 GB | Summaries and classifications |
| granite4.1:3b | 2.1 GB | Lightweight tasks, low latency |
| llama3.2:3b | 2.0 GB | Backup, multilingual support |
| bge-large:335m | 670 MB | Embeddings for Qdrant |
| qwen3-embedding:0.6b | 639 MB | Fast embeddings |
| nomic-embed-text | 274 MB | Embeddings fallback |
| smollm2:135m | 270 MB | Tests and trivial tasks |
The largest model (8B) fits in 16 GB of RAM without touching swap. Embeddings are quantized so semantic search returns results in under 100 ms.
The real cost: what you pay vs what you’d save
What d0a1.es pays per month
| Item | Cost / month |
|---|---|
| VPS Clouding 4×16 | ~€25 |
| Domain d0a1.es | ~€1 |
| Traffic, backups, misc | ~€2 |
| Local AI (Ollama + Qdrant + n8n) | €0 |
| Total | ~€28 / month |
What it would cost to replicate this stack with SaaS
Same traffic, same operations. One-to-one equivalent services:
| Equivalent SaaS service | Estimated cost / month |
|---|---|
| OpenAI API (generation + embeddings, ~50 posts/month) | ~€50 |
| Algolia or Pinecone (semantic search) | ~€30 |
| Cloudflare Pro + WP Engine (hosting) | ~€50 |
| Buffer / Hootsuite (social publishing) | ~€30 |
| Akismet (comment moderation) | ~€10 |
| BlogVault (backups) | ~€25 |
| ManageWP (multi-site management) | ~€25 |
| Total equivalent SaaS | ~€220 / month |
Direct savings: ~€192 / month, around €2,300 / year. And we haven’t counted the opportunity cost of SaaS price hikes, the migrations when they change their API, or the GDPR compliance cost of passing data to third-party servers.
Public repos (the “open source” proof)
14 repos at git.d0a1.es. The 6 public ones are the foundation of the catalog and the transparency claim:
d0a1/skills— 6+ working skills (image-alt-generator, link-validator, draft-reviewer, content-review, guardian, seo-content).d0a1/agentic-standards— operating standards for agents.d0a1/spec-driven-development— SDD flow we use for releases.d0a1/text-manipulation-tools— text utilities used by skills.d0a1/tpl-next-saas— Next.js 16 template that powers Storeroom OS.d0a1/crew-ops-platform— multi-agent orchestration.
What is NOT automated (brutal honesty)
If we tell you what works, we also tell you what doesn’t yet:
- ❌ “5 min deployment” — never formally measured. Reality is >30 min on a clean VPS.
- ❌ “100+ skills/extensions” — the catalog has 6 + 6 + 6. Inflated figure that needs correction.
- ❌ “Live demo” —
/demo/has a textual walkthrough, not an exposed stack. F2 of the roadmap. - ❌ Signed testimonials — this case is the first. No public customers yet.
- ❌ Disk at 92 % — 8.2 GB free. It works, but space needs freeing before something fails.
If you see something we claim elsewhere on this site that this case doesn’t back up, it’s a marketing bug to fix, not an excuse.
How to verify all of this
Any visitor can audit the current state. Here are the exact commands we used for this case:
# VPS state
ssh hermes@85.208.21.51 'uptime && df -h / && free -h'
# Active containers
ssh hermes@85.208.21.51 'sudo docker ps --format "table {{.Names}}\t{{.Status}}"'
# Available Ollama models
ssh hermes@85.208.21.51 'sudo docker exec ollama ollama list'
# Served subdomains
curl -sI https://d0a1.es/
curl -sI https://app.d0a1.es/
curl -sI https://git.d0a1.es/
curl -sI https://storeroom.d0a1.es/If you run these commands at any time, you’ll see the real state of the stack. This page is updated when the state changes significantly.
Want to do the same on your VPS?
The WordPress + AI Kit is the shortest path. It includes a deployment checklist, 15 operational prompts, and the cost comparison:
Audit performed on June 3, 2026, 21:02 UTC. Case #001 — self-documented. The meta-case.