What to Watch Tonight

Most recommendation tools either guess your taste badly or suggest titles that are not actually available anywhere you can watch them. What to Watch Tonight fixes both problems. It builds recommendations from either an enriched watch-history profile or a simple favorites list, then filters those results against current TMDb watch-provider availability so the shortlist is personal and immediately useful.

That makes it practical in the real-world moment that matters: deciding what to watch tonight without spending 45 minutes scrolling. Users can come in with a Netflix-style history profile, or just name 5 to 10 titles they love, and get movie or series picks that reflect actual taste instead of generic popularity.

The skill is built for speed and realism. It avoids already watched or seeded titles, supports both history-based and lightweight manual-input workflows, and works across real providers like Netflix, Disney+, Prime Video, and others supported through TMDb availability data.

Setup is lightweight but honest. Users need a free TMDb account plus either an API key or read token, and the package includes a step-by-step setup guide, example wording for the TMDb request form, and a fastest-path verification flow. The result is a small utility skill that does one job very well: help you stop browsing and start watching.