For music catalogs and sync teams

Find the right track by describing it.

Spondulix Search turns a plain-English request into a precise, ranked search across your catalog. Genre, mood, tempo, lyrics, era, sounds-like: ask for it the way you'd ask a colleague.

Get started See how it works

Works with your existing catalog data. Web interface for editing, REST API for everything else.

Built for how people actually look for music

Everything a search needs to understand a brief, not just match keywords.

Natural language queries

Describe what you need in everyday words. "Something like early Coldplay, hopeful, no drums until the chorus" is a valid search.

Precise filters

Combine free text with hard constraints: tempo ranges, duration, instrumental only, clean versions, public domain, your own custom tags.

Weighted results

Every part of the request carries a weight, so the tracks that satisfy the most important parts rise to the top instead of a flat keyword match.

Semantic lyric search

Search lyrics by theme, not exact words. "Songs about leaving home" finds them whether or not the word "home" appears.

REST API and MCP

Load your catalog, manage metadata and run searches from your own systems. AI assistants like Claude can connect directly through the MCP server.

Multiple libraries

Keep catalogs separate, invite collaborators with the right level of access, and bring your own AI provider keys per library.

How it works

Three steps from a spreadsheet of tracks to a searchable catalog.

1

Load your catalog

Add tracks, composers, publishers and your own metadata through the web interface or the API. Lyrics are indexed automatically.

2

Tune the prompt

Each library has an editable prompt and examples that teach the search how your catalog is tagged and what your users ask for.

3

Start searching

Type a brief, get a ranked list. Refine it in conversation: "the same, but slower" continues the search rather than starting over.

Ask it like you'd ask a music supervisor

A few of the kinds of requests it handles.

"Cinematic strings that build for 30 seconds then drop out" "90s hip hop, clean version, under three minutes" "Acoustic instrumental, wistful, for a documentary about farming" "Something that sounds like Fleetwood Mac but public domain" "Lyrics about second chances, male vocal, mid tempo" "Same as the last search but faster and no vocals"
For developers

One request, ranked results

The same search that powers the web interface is a single JSON call. Send a natural language string, optional structured filters and a sort order, and get back scored tracks from your own catalog.

  • Bearer-token authentication, plain JSON in and out
  • Full CRUD for tracks, composers, publishers and metadata
  • Continue a search to refine it, the way a conversation would
  • Paginated lists with predictable, documented shapes
Read the API docs
POST /api/v1/libraries/{library_id}/searches 
curl -X POST https://app.spondulix.ai/api/v1/libraries/42/searches \
  -H "Authorization: Bearer $SPONDULIX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "search_params": {
      "nls": "wistful acoustic instrumental, under three minutes",
      "filter": { "and": [
        { "instrumental": { "value": true } },
        { "duration": { "less_than": 180 } }
      ]},
      "sort": [["total_score", "desc"]]
    }
  }'

Plans sized to your catalog

Every catalog is different, so pricing is too. Whether you're an independent label with a few hundred tracks or a publisher with a million, we'll put together a plan that fits, and you can bring your own AI provider account.

Get in touch