Skip to main content
Trackyard’s search engine understands context, not just keywords. Describe what you need in plain English, and the AI infers genre, mood, BPM, instrumentation, and more.

​
How It Works

Traditional music search requires you to know exactly what filters to use. Trackyard’s AI interprets your intent and translates it into structured queries behind the scenes.
Query: “happy upbeat music”What happens:
  • Searches for tracks tagged with “happy” or “upbeat”
  • Returns thousands of results
  • No understanding of context or use case
  • You spend 20 minutes filtering by BPM, genre, vocals, etc.
Result: Overwhelming, time-consuming, often misses the right track.

​
Natural Language Understanding

The AI search engine understands:
Examples:
  • “moody piano for a rainy scene”
  • “chill lo-fi for a coffee shop vlog”
  • “dramatic orchestral for a movie trailer”
What the AI infers:
  • Instrumentation (piano, lo-fi beats, orchestra)
  • Mood (moody, chill, dramatic)
  • Energy level (low, medium, high)
  • Genre (classical, hip hop, cinematic)
Examples:
  • “background music for a real estate walkthrough”
  • “15-second clip for an Instagram Reel”
  • “podcast intro music”
What the AI infers:
  • Duration requirements (15 seconds, ~1 minute, ~30 seconds)
  • Energy level (low/ambient for real estate, higher for reels)
  • Vocal preference (instrumental for background, vocal hooks for intros)
  • Mood (spacious, upbeat, professional)
Examples:
  • “fast-paced electronic without vocals”
  • “slow acoustic guitar in D minor”
  • “130 BPM trap beat”
What the AI infers:
  • BPM (fast = 140+, slow = 60-90, specific = 130)
  • Vocal presence (without vocals = instrumental only)
  • Key signature (D minor)
  • Genre (electronic, acoustic, trap)
Examples:
  • “something warm and nostalgic”
  • “tense and suspenseful”
  • “bright and optimistic”
What the AI infers:
  • Mood tags (warm, nostalgic, tense, suspenseful, bright, optimistic)
  • Instrumentation (warm = acoustic/analog, tense = strings/bass)
  • Energy level (suspenseful = mid-high, optimistic = high)

​
How to Write Effective Queries

1

Be specific about the use case

✅ Good: “upbeat music for a tech product demo”❌ Bad: “upbeat music”Why it matters: Context helps the AI infer mood, energy, and instrumentation.
2

Include mood or vibe descriptors

✅ Good: “moody lo-fi piano for a rainy scene”❌ Bad: “lofi piano”Why it matters: Mood descriptors help the AI understand emotional tone and tempo.
3

Specify technical requirements if needed

✅ Good: “130 BPM trap beat without vocals”❌ Bad: “trap beat”Why it matters: If you have specific technical needs (BPM, key, vocals), include them upfront.
4

Use reference tracks or scenes

✅ Good: “something like the music in that Apple commercial”✅ Good: “similar to Tycho but more upbeat”Why it matters: References give the AI a starting point for instrumentation and production style.

​
Combining AI Search with Filters

For maximum precision, combine natural language queries with structured filters: 
{
  "query": "upbeat music for a tech startup video",
  "filters": {
    "has_vocals": false,
    "min_bpm": 120,
    "max_bpm": 140,
    "energy_level": "high",
    "genres": ["Electronic", "Pop"]
  }
}
The AI interprets the query, then applies the filters on top for surgical precision.

​
What the AI Can’t Do (Yet)

Reference track uploads: Future feature. Currently, you can describe a reference track in text (“something like Tycho”), but you can’t upload an audio file for similarity matching.
Emotion detection from video: The AI doesn’t analyze video content directly. Describe the scene or emotion in your query instead.

​
Search Tips & Examples

Query: “15-second upbeat clip for an Instagram Reel about travel”Why it works: Duration + vibe + use case gives the AI everything it needs.
Query: “chill lo-fi background music for a coding tutorial”Why it works: “Background music” signals low energy + instrumental. “Coding tutorial” reinforces focus/productivity vibe.
Query: “short upbeat intro music for a business podcast, 20-30 seconds”Why it works: Duration + energy level + genre context (business = professional, clean).
Query: “minimal electronic music for a SaaS product demo, modern and clean”Why it works: “Minimal” + “modern and clean” infers instrumentation (synths, sparse drums, no vocals).
Query: “tense orchestral music for a chase scene, building intensity”Why it works: Mood (tense) + genre (orchestral) + scene type (chase) + dynamic (building).

​
Advanced Features (Paid Plans)

​
Intent Extraction

On paid plans, the AI automatically extracts structured intent from your query and displays it: Query: “moody piano for a rainy scene” Extracted Intent: 
{
  "genres": ["Classical", "Ambient"],
  "moods": ["Moody", "Melancholic", "Atmospheric"],
  "instruments": ["Piano"],
  "energy_level": "low",
  "min_bpm": 60,
  "max_bpm": 90,
  "has_vocals": false
}
You can edit the extracted intent on the fly to fine-tune results.

​
Next Steps

Smart Clip Trimming

Learn how automatic segment selection works

Search API Reference

See all available search parameters

API Quickstart

Try your first search

Use Cases

See AI search in action