Gogocho Cinema

Choosing a film is easy until everyone has a different mood and a hard time to get home. Gogocho turns those constraints into a conversation with a warm, fictional cinema host. Ask for something light tomorrow, under two hours and finishing before ten; the programme responds on screen. Compare two films, choose a screening and ask for two tickets. The agent prepares a review, while the visitor makes the final confirmation. AssemblyAI’s Voice Agent API powers live speech recognition, conversational reasoning, spoken replies and tool calling. A browser AudioWorklet converts microphone audio to 24 kHz PCM. Three tools search screenings, compare films and prepare bookings. A small Node.js backend validates dates, strict finish times, film identifiers and quantities, and calculates prices. The model cannot confirm a booking or invent an available screening. The minimal interface includes original illustrated posters, visible conversation status, transcripts, microphone selection and a local microphone test. Visitors can browse manually or try a clearly labelled scripted preview without a microphone or API session. Confirmed demo tickets can be saved to a calendar with the fictional status retained. All eight films, programme entries, prices and bookings are original and fictional. There is no payment, real seat inventory, external film scraping or third-party film licensing dependency. The project is a solo, dependency-free JavaScript prototype designed to demonstrate a useful voice-to-interface workflow with a clear human confirmation boundary.