Blog
Notes from the platform.
Updates on Cadesia Pod, the app library, and how the platform is built.
9Router vs OpenRouter: self-hosted or hosted AI routing
OpenRouter sells credits for 300+ models with a 5.5% purchase fee. 9Router is self-hosted and routes your own keys and subscriptions. Which fits you.
Run an always-on personal AI agent on Telegram with Hermes
An agent that answers on Telegram, remembers you, and runs daily briefings on a schedule. What it takes to keep Hermes Agent running 24/7.
Cadesia Pod vs a VPS or Railway for self-hosted apps
Running n8n, WAHA or Uptime Kuma on a VPS, on Railway, or in a Cadesia Pod: what each costs, what you still have to manage, and when each makes sense.
Hermes Agent vs ChatGPT: run your own agent or rent one?
ChatGPT is a polished assistant you rent. Hermes Agent is an open-source agent you run: any model, your channels, your data. How to choose.
Self-hosted n8n vs n8n Cloud: what you really pay for
n8n Cloud starts at €24 a month for 2,500 executions. Self-hosted n8n is free, but you run the server. The real costs of both, and a third option.
n8n vs Make: credits, code, and who holds your data
Make bills credits per module run and hosts everything. n8n counts whole workflow runs, or none at all when self-hosted. Where each one wins.
One API key for every AI model: route them with 9Router
Give Claude Code, Cursor, n8n and your agents one OpenAI-compatible endpoint and key, and let 9Router choose the provider and fall back when one runs out.
WAHA vs WhatsApp Cloud API: which to build on
Meta's official Cloud API charges per template message and needs approved templates. WAHA links a normal number like WhatsApp Web. Here's when each fits.
Add a public status page to your website with Uptime Kuma
Set up monitors, publish a public status page, and get alerts on Telegram or Slack. A self-hosted status page with Uptime Kuma in about ten minutes.
WhatsApp auto-replies with WAHA and n8n, no backend code
Connect a WhatsApp number to n8n with WAHA and reply to messages automatically: webhook in, logic in the middle, sendText out. Step by step.
Solutions are live: tell us the outcome, we pick the Apps
Solutions turn an outcome into ready-made Apps — one subscription, a bundle price, and every App still in its own Pod. Here is why we built them.
AI news: science, adoption, and safety
Three AI stories from September 2026: Claude's biology research, Google's AI adoption data, and OpenAI's call for safety rules.
TAI-DR and the meat proxy in your pull request
An AI agent writes the patch, a developer forwards it, and a reviewer does the first real reading. How TAI-DR and the meat proxy describe the same failure.
What is 9Router? One endpoint for every AI provider
9Router is a self-hosted AI router: one OpenAI-compatible endpoint for 40+ providers, with token savers, quota tracking and automatic fallback.
n8n webhooks: trigger a workflow from anything that can POST
How the n8n Webhook node works — test vs production URLs, response modes, verified signatures, and a first webhook in minutes.
n8n 2.40 is out — AI fallback models, and the 2.x features worth trying
n8n 2.40.0 lets AI nodes fall back to a second model, plus what else landed across recent 2.x releases — and how to run it without a server.
Cadesia Pod is live
Cadesia Pod is live at console.cadesia.com: pick an App like n8n, WAHA or Uptime Kuma and get an isolated Pod with resources tuned for the workload.
Uptime Kuma: self-hosted status monitoring in minutes
What Uptime Kuma monitors, how monitors, status pages, and alerts work — and the fastest way to run your own instance.
Pick an app, not a server
Why Cadesia starts from workloads instead of servers — and what we took away so you don't have to think about it.
Hermes Agent: run your own always-on AI agent
What an always-on agent runtime actually does — memory, skills, channels, and a dashboard — and why it needs a dedicated home.
Send WhatsApp messages from your app with WAHA's REST API
WAHA gives WhatsApp a developer interface — sessions, REST calls, and webhooks. Here's how the pieces fit and how to run it.
n8n vs Zapier: when to self-host your automations
Per-task pricing vs a flat instance, data that stays with you, and where self-hosting n8n actually wins — plus the one tradeoff to solve.
What a Pod gets
Isolation, an Address, persistent storage, and managed runtime — the four things every Pod ships with.
UptimeRobot vs Uptime Kuma: hosted SaaS or your own monitor
Monitor caps, status pages, alerts, and pricing — how hosted uptime monitors compare to self-hosted Uptime Kuma.
