DEV Community

Cover image for Aition AI
Artem
Artem

Posted on

Aition AI

Notion MCP Challenge Submission 🧠

Aition AI β€” The Autonomous Executive AI CTO for Notion 🧠

The Project Tagline
Bridge the gap between raw code and strategic engineering intelligence. Aition is an autonomous AI CTO that lives in your Notion.

Inspiration: The Hidden Engineering Tax
As a developer, I noticed that engineering teams lose up to 40% of their velocity to "context switching" and technical debt. Developers spend hours manually updating tasks, and managers have no real-time visibility into codebase health.

I built Aition AI to kill this overhead. It's a "human-in-the-loop" system that monitors GitHub activity and transforms it into high-level executive intelligence inside Notion.

What is Aition AI?
Aition acts as your team's central nervous system. It’s an autonomous platform that:

Audits every commit using LLMs (Claude 3.5/GPT-4o) via OpenRouter.

Quantifies technical debt and assigns a "Health Score" to every push.

Generates unique NFT identities for developers using a custom pixel-art engine.

Synchronizes everything to Notion, creating a command center that stays updated without anyone typing a single word.

The "Superpower" Workflow
Developer pushes code β†’ GitHub Webhook fires β†’ Aition AI Audits β†’ Notion DB updates.

Critical bugs (like auth crash or hotfix) trigger instant P0 incident pages in Notion.

No more manual standup reports. No more "forgotten" tech debt.

The Architecture & Tech Stack
I designed Aition as a robust monorepo capable of handling global event streams.

Frontend: Next.js 14 (App Router) with a premium "Vercel-style" dashboard.

Backend: Node.js & Express (Hosted on Render) orchestrating the AI logic.

AI Core: OpenRouter API (Accessing Claude 3.5 Sonnet for deep code reasoning).

Database: Supabase (PostgreSQL) for team metadata and historical analytics.

Output Layer: Notion API & MCP-inspired service for database orchestration.

How I used Notion MCP
Notion is the brain of this system. I designed a 4-database ecosystem to manage the engineering lifecycle:

Audits Database: Logs every technical check with AI-generated summaries and scores.

Tasks Database: Automatically converts commits into trackable tasks with AI-assigned priorities.

Incidents Database: A high-severity hub for critical code risks.

Team Identity DB: Stores developer profiles and their unique generated NFT avatars.

Challenges I Overcame
Strict Data Validation: Ensuring AI-generated JSON perfectly matches Notion’s database properties was a challenge. I built a validation layer to prevent API errors.

Async Reliability: GitHub webhooks require immediate responses. I implemented an asynchronous workflow where the server confirms receipt instantly while the AI analysis runs in the background.

Identity Generation: Building a pixel-art engine that generates unique avatars from name hashes was a fun challenge in algorithmic creativity.

What I Learned
Building Aition taught me that the best DevOps tools are the ones that work when you're not watching. I mastered the discipline of secure API management and learned how to leverage Notion not just as a doc, but as a fully functional execution engine.

The Repository & Demo
GitHub: https://github.com/vanzero1448/AiTion-Your-bussines-without-coding

Live Demo: https://a-ition-fro.vercel.app

Video you can download here - https://wdho.ru/cjS9

License: MIT

Built by vanzero1448 for the Notion MCP Challenge 2026.

Top comments (0)