Frontier AI. Live web. Real insight.

Build "For You"
that actually
understands why.

TasteRay's Emotional AI doesn't just match preferences. It understands the deeper motivations driving your users' choices across any vertical.

<3s
P95 Latency
99.9%
Uptime SLA
25+
Verticals
api.tasteray.com
Request
{ "vertical": "products", "context": { "profile": "Looking for premium wireless earbuds for daily commute. Noise cancelling." } }
Response
Sony WF-1000XM5 96% match
Why you'll love it "You're not buying earbuds—you're buying back your mental space."
Key factors
  • 30dB cancellation makes the subway disappear
  • 8-hour battery outlasts your longest day
  • $279 — flagship quality, room to spare

Generic recommendations
kill engagement.

Traditional recommendation engines see patterns. They don't understand context, mood, or the real reasons behind user choices.

😭

Cold Start Paralysis

New users get generic suggestions that feel impersonal. They bounce before your algorithm can learn.

📈

Shallow Matching

Collaborative filtering shows "similar users liked this" but can't explain why it matters to this specific user.

Stale Context

Batch-processed models miss real-time signals. Your user's mood changed, but your recs didn't.

Emotional AI that understands
the why behind every choice.

TasteRay combines frontier LLMs' reasoning with real-time web grounding to deliver recommendations that feel genuinely personal and explainable.

🧠

Deep Understanding

Pass unstructured user profiles, conversation excerpts, or simple preference lists. TasteRay understands context in any format and identifies emotional drivers.

🌐

Real-time Grounding

Every recommendation is grounded with live web data. New releases, updated ratings, current availability - always fresh, always relevant.

💬

Explainable Matches

Every recommendation comes with human-readable explanations. Tell users exactly why they'll love it, building trust and boosting engagement.

🎯

Any Vertical, One API

Movies, restaurants, products, travel, jobs, real estate - one unified API handles 25+ verticals with the same simple interface.

<3s
P95 Query Latency
99.9%
Uptime SLA
25+
Supported Verticals
50K+
Recommendations Served

Everything you need for
world-class personalization.

A complete recommendation infrastructure that scales from prototype to production.

🔍

Web Grounding

Real-time search via Exa provides current information - new releases, reviews, availability. No stale data.

📝

Flexible Input

Structured arrays or natural language profiles. Pass conversation excerpts, survey responses, or simple lists.

🎓

Rich Explanations

Every recommendation includes why_match, key_factors, and potential_concerns. Build transparent "For You" experiences.

Fast Mode

Enable fast inference for time-sensitive applications. Same quality, lower latency for real-time scenarios.

🌍

Multi-language

Generate recommendations and explanations in any language. Just set the locale code in your request.

📊

Usage Analytics

Track requests by vertical, monitor rate limits, and analyze patterns with the built-in usage endpoint.

One API. Every vertical.

From entertainment to e-commerce, TasteRay adapts to your domain without custom model training.

🎬 Movies
📺 TV Series
📖 Books
🎵 Music
🎙 Podcasts
🎮 Games
🍽 Restaurants
🏨 Hotels
🛫 Flights
🏝 Destinations
🏃 Activities
🎟 Events
💻 Electronics
👕 Clothing
📚 Courses
💼 Jobs
🏠 Real Estate
🚗 Cars
Coming Soon

Your products.
Our understanding.

One URL. Infinite personalization.

Point TasteRay at any product feed—XML, JSON, CSV, Google Shopping, Shopify, whatever format you have. We detect it automatically, index every product, and keep it synced in real-time.

No configuration. No schema mapping. No manual setup. Your recommendations now draw from your actual inventory. Updated prices. Real availability. Products your customers can actually buy.

You provide
{ "feed_url": "shop.co/products.xml" }
We deliver
👟 Nike Air Max 90 94%
👔 Patagonia Fleece 91%
Garmin Instinct 2 87%
Coming Soon

Every click
teaches.

Your recommendations get smarter. Automatically.

Send us what your users actually do - views, clicks, purchases, dismissals. We turn that signal into sharper recommendations without you touching a single model.

No ML expertise required. No training pipelines to maintain. Just a simple feedback endpoint that makes every recommendation better than the last.

Coming Soon

npm install
done.

TypeScript. Python. Whatever you ship with.

Type-safe clients that make integration a one-liner. Full autocomplete, proper error handling, and zero boilerplate for recommendations and feedback.

Plus framework starters for Next.js and Cloudflare Workers. Drop in a component, get a "For You" feed with SSR and caching built in.

TypeScript / Node
npm install @tasteray/sdk
Soon
Python
pip install tasteray
Soon
Next.js Starter
npx create-tasteray-app
Soon
Coming Soon

Memory that
matters.

One user. Infinite context.

Persistent taste profiles that grow with every interaction. Explicit preferences merge with implicit signals to build a rich understanding of each user over time.

The longer someone uses your product, the more personal their recommendations become. No cold starts. No amnesia. Just recommendations that feel like they know you.

👤
Sarah's Taste Profile
usr_8x7f2k9m4n...
Adventure
0.85
Value-driven
0.72
Eco-conscious
0.91
Minimalist
0.64
Coming Soon

Prove
it.

See exactly what works.

Built-in A/B testing that tells you whether TasteRay is actually moving the needle. CTR, conversion, engagement - tracked automatically with statistical significance.

Tag requests with experiment IDs. Compare variants. Get clear answers without a data science degree. Because "it feels better" isn't a business case.

Homepage Recs Test Significant
Control Baseline
CTR 3.2%
Conversion 1.8%
Revenue/user $12.40
TasteRay Winner
CTR 5.1%
Conversion 2.9%
Revenue/user $18.20

Ship in minutes,
not months.

A simple REST API with comprehensive documentation. No ML expertise required.

  • RESTful JSON API - works with any stack
  • Detailed API documentation with examples
  • Test keys for development, live keys for production
  • Webhook support for async workflows (coming soon)
  • Rate limiting with clear headers and backoff guidance
Read the Docs
// Get personalized recommendations
const response = await fetch('https://api.tasteray.com/v1/recommend', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'X-API-Key': process.env.TASTERAY_KEY
  },
  body: JSON.stringify({
    vertical: 'products',
    context: {
      profile: 'Looking for noise-cancelling
        earbuds for daily commute.',
      constraints: { budget_max: 300 }
    },
    options: { count: 5 }
  })
});

const { recommendations } = await response.json();
// recommendations[0].item.name = "Sony WF-1000XM5"
// recommendations[0].explanation.why_match = "..."
# Get personalized recommendations
import requests
import os

response = requests.post(
    'https://api.tasteray.com/v1/recommend',
    headers={
        'Content-Type': 'application/json',
        'X-API-Key': os.environ['TASTERAY_KEY']
    },
    json={
        'vertical': 'products',
        'context': {
            'profile': 'Looking for noise-cancelling
                earbuds for daily commute.',
            'constraints': {'budget_max': 300}
        },
        'options': {'count': 5}
    }
)

recs = response.json()['recommendations']
# recs[0]['item']['name'] = "Sony WF-1000XM5"
# Get personalized recommendations
curl -X POST https://api.tasteray.com/v1/recommend   -H "Content-Type: application/json"   -H "X-API-Key: $TASTERAY_KEY"   -d '{
    "vertical": "products",
    "context": {
      "profile": "Looking for noise-cancelling
        earbuds for daily commute.",
      "constraints": {"budget_max": 300}
    },
    "options": {"count": 5}
  }'

# Response: {"recommendations": [...]}

Start free. Scale as you grow.

No credit card required. Upgrade when you're ready.

Free
$0/mo
For prototyping and testing
  • 1,000 fast requests/month
  • 5 requests/minute
  • All verticals
  • Community support
Get Started
Basic
$49/mo
For small apps and MVPs
  • 10,000 fast requests/month
  • 50 requests/minute
  • All verticals
  • Email support
Start Trial
Enterprise
Custom
For scale and compliance
  • Unlimited requests
  • Custom rate limits
  • SLA guarantee
  • Dedicated support
Contact Sales

Ready to build "For You"
that actually works?

Join teams shipping personalized experiences that users love. Start with 1,000 free requests.