Build an app to uncover fun facts from your Netflix viewing history, with OpenAI, Vercel & Gandalf SDK
whoami.tv is an open source app that shows you fun facts about yourself based on your Netflix viewing history.
Try it out yourself & then check out the code on GitHub.
The Stack
Infra
whoami.tv is built with Next.js & Vercel. It uses a lot of stuff from Vercel — Vercel Postgres, KV (Redis), Functions & an integration with Inngest for managing background jobs.
Building everything with Vercel’s vast product offering allowed us to focus entirely on the application without having to worry about infrastructure management or resource provisioning.
Additionally, Vercel makes it super easy for you to spin up your own instance of whoami.tv, using the Deploy Button.
Getting Netflix Data
Gandalf is an SDK for securely requesting user data from services like Netflix, Amazon, Uber etc. It unlocks for a whole new class of applications that couldn’t have been built before as these services did not have a first party API.
See how it works:
AI & LLMs
whoami.tv uses the OpenAI Assistants & Perplexity APIs to:
- Generate witty quips about your stats or fun facts. This makes the application feel truly personalised to the end-user. Combining LLMs & data from Gandalf is an arena that gets us really excited.
- Perplexity has “online models” that can provide up-to-date, and factual responses. This was useful for programmatically getting the personalities of characters in TV as part of the process of determining your TV BFF
You can dive into all the AI assistants code on Github.
If you have any feedback, issues playing around with this code, or just questions in general, please send them my way.
Or join the Gandalf Developer Slack.
