mymem SabaKan

Inspiration

Server operations are fragmented across monitoring tools, terminals, documentation, and individual engineers' memories. MyMem SabaKan connects these activities into one traceable workflow.

What it does

MyMem SabaKan is an AI-guided DevOps workspace that:

• Monitors servers and WordPress environments
• Detects HTTP and service incidents
• Executes SSH commands across multiple servers
• Analyzes logs and command results with GPT-5.6
• Manages Incus environments and ZFS snapshots
• Preserves operational decisions and execution history in MyMem

How we built it

The project uses Codex and GPT-5.6 together with TypeScript, React, Node.js, Incus, ZFS, SSH, nginx, PHP-FPM, MariaDB, WordPress, DuckDB, WebSocket, Cloudflare, and ConoHa VPS.

Codex was used for architecture, implementation, debugging, UI development, and documentation. GPT-5.6 interprets server status, proposes investigation procedures, analyzes execution results, and turns them into reusable operational knowledge.

Challenges

One major challenge was correctly separating operating-system status from web-service status. A server may still be running while PHP-FPM is stopped and nginx returns a 502 error.

Accomplishments

A workflow that previously required approximately three hours can now be completed in approximately thirty minutes, while improving consistency and preserving the reasoning behind each operation.

What's next

We plan to add safer approval controls, environment cloning, deployment workflows, incident timelines, shared operational playbooks, and additional MCP integrations.

Built With

• chrome
• cloudflare
• codex
• conoha-vps
• duckdb
• gpt-5.6
• incus
• node.js
• openai
• react
• ssh
• typescript
• websocket
• wordpress
• zfs