Articles and courses about MCP, AI integration, and software development
Welcome to SFLOW. Here we share insights, tutorials, and articles about topics we find interesting and useful. From AI integration and MCP servers to software development best practices. We hope you find something valuable.
Get dramatically better results from AI tools you already use. Better prompting, Claude Projects, document analysis — no coding required.
Start Learning →Transform from ChatGPT user to someone who can actually build and deploy websites, tools, and automations with AI assistance.
Start Building →Autonomous AI agents need infrastructure: isolated, persistent workspaces where they can run unattended without putting your other clients' data at risk. How we built a self-hosted agent server with Coder, Claude Code, and Playwright — and why the whole industry is converging on the same pattern.
A model has a hard time catching its own class of mistake. So we have one AI write, and two others try to break it — Claude Code shells out to OpenAI's Codex and Google's Gemini, then fact-checks every finding against the real code. Here is the exact setup, what it caught, and where it bites.
We pointed Claude at EnvisionSCADA's public demo and watched it configure a power meter, build a 30-widget dashboard, write a simulation script, and verify everything in a live browser. The lesson is bigger than the demo: web-first, open platforms are where agentic engineering happens.
How we replaced a no-code AppSheet solution with a full-stack Next.js platform for a European textile manufacturer — using Claude Code as a development accelerator.
We used Claude Code with 20 parallel AI agents to research 90 companies across three European stock indices, generate personalized outreach, and deploy a live CRM — all from one VS Code terminal.
We triggered a cascading pump failure on a live Ignition gateway and asked Claude to find the root cause. Using MCP tools, it traced 6 alarms across 3 process areas to a single equipment failure in under 90 seconds.
How we built a 66-tool MCP server that gives Claude live access to Ignition SCADA gateways — including a self-deploying API that builds and installs 20 WebDev endpoints in-memory with a single REST call.
How we built a 75-command CLI for Inductive Automation's Ignition 8.3 REST API with Claude Code — and why deterministic tooling still matters in the age of AI agents.
Open any folder, choose your AI model, and let agents work in the background. VS Code with terminal AI agents, MCP servers, and the extension ecosystem is the AI workbench most teams are already sitting on.
How I built a scheduling platform for Claude AI that lets you configure recurring tasks and event-driven workflows with natural language - moving beyond question-answer AI.
What if you could plan your Belgian train journeys, check your CRM, and draft emails - all from one AI conversation?
How MCP Servers act as the USB-C of AI - one plug that works across all your business systems.
