AI Portfolio Builder: An App for Developers Who Want a Portfolio Site in Minutes
Overview
Building my portfolio took me longer than most of the projects in it. And most developer portfolios end up looking the same: a dark page, a gradient, a grid of cards. So I took this site, cleaned it up and turned it into an open source template called AI Portfolio Builder, so anyone can get the same kind of site in a few minutes instead of a few weeks.
What you get
It is a complete portfolio website: a home page, selected work, project pages built from content blocks, side projects, an experience timeline, an FAQ, a CV download and a password protected admin panel to edit all of it. It works on phones, tablets and desktops. It ships with a fictional sample person, Odile, so you can see everything working before you add your own content.
Looks right on every screen
Every page has a phone and tablet layout, with a mobile menu, and the admin panel works on a phone too, so you can fix a typo or add a project from wherever you are.
Everything is ready
The parts that usually take the longest are already done: a database that sets itself up on the first deploy, Postgres with pgvector for the AI search, API routes and server actions, a protected admin panel, page metadata and generated share images, per visitor rate limits on the AI endpoints, and feature flags to switch any part off.
Three ways to set it up
The quickest way is the Deploy with Vercel button. It copies the repo to your GitHub account, creates a free Neon database, asks for three values and deploys. The first build creates the tables, loads the sample content and builds the search index by itself, so there are no commands to run.
On your own machine it is one command, npm run setup, which asks for your name, links, colour theme, database and AI key, writes the config for you and starts the site. The third way is to open the folder in your coding agent, such as Claude Code or Codex, paste the prompt from the README, and let the agent ask you the questions and set everything up.
Start from your LinkedIn, CV and GitHub
In all three ways you can give it your LinkedIn profile link, your CV and your GitHub username. Your own AI drafts the site from them: your profile, the home page text, your experience, FAQs and a project for each of your repositories. It runs on your own AI key or your own coding agent, never mine. You see a summary first and nothing is written until you approve it. It is told not to invent anything, and links or numbers that are not in your sources are dropped and listed. LinkedIn is read on a best effort basis: if your profile cannot be read, it tells you and asks for the PDF that LinkedIn's Save to PDF gives you.
The AI layer is optional
On top of the portfolio there is the same AI layer this site has: a chat that answers visitors' questions about your work only from your own content, a job fit check where a recruiter pastes a job description and gets an honest verdict, and an MCP server so other AI assistants can ask about you. It works with OpenAI by default, or with Claude, Gemini, OpenRouter or a local model through Ollama. Every part can be switched off, and without an AI key you simply get a plain portfolio.
Manage it your way
You can keep everything up to date in the admin panel in your browser, or just ask your coding agent: it adds or removes a project with one JSON file through the project scripts, the same way I add work to this site. After every change the chat's search index updates itself, with nothing to rebuild.
Why neo-brutalism
I chose neo-brutalism, with thick outlines, hard shadows and flat bold colours, because it still stands out wherever it is used and people actually remember it. It comes with ready made colour themes, including pastels and neutrals, switched with one line. If someone wants a completely different look, they can give their coding agent a design they like and ask it to restyle the site.
Open source
AI Portfolio Builder is free and open source under the MIT licence, on GitHub at mutlumehmet/ai-portfolio-builder. It is built with Next.js 16, Prisma 7 and Postgres with pgvector. To try it on your own machine:
git clone https://github.com/mutlumehmet/ai-portfolio-builder.git my-portfolio
cd my-portfolio && npm install && npm run setup