Streaming Recommendations

Recommend 3-5 movies or TV shows to watch on streaming services, tailored to the user's genre, mood, format preference, and available platforms.

When to Use

• User asks "what should I watch?"
• User wants movie or TV show recommendations
• User mentions streaming services and wants suggestions
• User describes a mood or genre and wants entertainment picks

Process

Step 1: Parse the Request

Extract any preferences the user already provided. Look for these four dimensions:

Step 2: Ask What's Missing

If any of the four dimensions are missing, ask one multiple-choice question at a time in this priority order. Skip any dimension already provided:

1. Format — "Are you in the mood for a movie or a TV show?"
   • Movie
   • TV show
   • Either is fine
2. Genre — "What genre sounds good?"
   • Action / Adventure
   • Comedy
   • Drama
   • Sci-Fi / Fantasy
   • Horror / Thriller
   • Documentary
   • Surprise me
3. Streaming service — "Which service are you watching on?"
   • Netflix
   • Hulu / Disney+
   • HBO Max
   • Prime Video
   • Apple TV+
   • Peacock / Paramount+
   • Any / I have them all
4. Mood — "What vibe are you going for?"
   • Light & fun
   • Intense & gripping
   • Thought-provoking
   • Heartwarming
   • Dark & gritty
   • Surprise me

If the user provides all four upfront (e.g., "recommend a funny movie on Netflix, something light"), skip straight to Step 3.

"Surprise me" on genre or mood means pick freely — lean toward highly-rated or trending picks.

Step 3: Search for Recommendations

Always use web search to find current recommendations. Streaming catalogs change constantly — never rely solely on training data.

Build search queries combining the user's filters:

• Example: "best light comedy movies on Netflix 2026"
• Example: "top intense sci-fi TV shows on HBO Max 2026"

Search 2-3 queries to cross-reference availability and find quality picks. Prioritize:

• Currently available on the specified service (verify availability)
• Well-rated (IMDb 7.0+ or Rotten Tomatoes 75%+)
• Mix of well-known and lesser-known titles for variety
• Matches the stated mood/vibe, not just the genre

Step 4: Present as Rich Cards

Return 3-5 recommendations using this format for each:

### {emoji} Title (Year)
**Genre:** Genre1, Genre2 | **Rating:** Score (Source)
**Streaming on:** Service Name

Synopsis in 2-3 sentences. Spoiler-free. Written to sell the vibe, not just
summarize the plot. Make it compelling.

**Why you'll like it:** One sentence connecting this pick back to the user's
stated mood and preferences.

Formatting rules:

• Use 🎬 for movies, 📺 for TV shows
• Rating from IMDb (e.g., "8.2 IMDb") or Rotten Tomatoes (e.g., "94% RT") — whichever the search surfaces
• Synopsis: 2-3 sentences, spoiler-free, emphasize tone and vibe over plot details
• "Why you'll like it": Directly reference the user's mood or preferences — this is what makes it feel personal

Example Output

You asked for: An intense sci-fi TV show on Netflix

📺 3 Body Problem (2024)

Genre: Sci-Fi, Mystery | Rating: 7.7 IMDb Streaming on: Netflix

A mystery that spans decades and continents as a group of brilliant scientists confront the greatest threat in humanity's history. The stakes escalate relentlessly, blending real physics with jaw-dropping sci-fi concepts.

Why you'll like it: You wanted intense and gripping — this one doesn't let you breathe from the first episode.

---

📺 Silo (2023)

Genre: Sci-Fi, Drama | Rating: 8.0 IMDb Streaming on: Apple TV+

In a ruined and toxic future, thousands live in a giant underground silo. When the sheriff investigates what's outside, she uncovers shocking truths that could destroy the community.

Why you'll like it: Slow-burn intensity that builds and builds — perfect for that edge-of-your-seat craving.

Key Principles

• Always search the web — streaming catalogs change; never guess availability
• One question at a time — don't overwhelm with multiple questions
• Personalize the "why" — tie every recommendation back to what the user asked for
• Spoiler-free always — never reveal plot twists or endings
• Variety matters — mix popular and under-the-radar picks
• Verify availability — if unsure a title is on the stated service, say so
• One and done — present the picks and let the user take it from there; don't prompt for follow-up