Gigoo

Inspiration

I wanted to make freelance marketplaces easier for both clients and freelancers by letting AI agents work with structured marketplace data instead of scraping the UI.

What I built

Gigoo is a Rails marketplace with eight WebMCP tools for searching jobs and freelancers, checking reputation, managing shortlists, and creating proposal drafts.

Clients can find trusted freelancers, while freelancers can discover suitable jobs and prepare proposals.

How I built it

Gigoo uses Ruby on Rails, PostgreSQL, Hotwire, Stimulus, Tailwind CSS, and WebMCP.

The WebMCP tools use the user's existing authenticated session and Rails authorization rules.

Challenges and learning

The main challenge was deciding what agents should be allowed to do. Gigoo lets agents search, evaluate, shortlist, and create drafts, while important actions like submitting a proposal remain human-controlled.

I learned that WebMCP works best when it exposes real product capabilities instead of adding a separate AI chat layer.

Built With

• hotwire
• javascript
• postgresql
• render
• ruby-on-rails
• stimulus
• tailwind-css
• turbo
• webmcp