🔎 Search & Recommendations

Search should help users find what they already know. Recommendations should help them keep going once they have started. Tuned Global’s Metadata APIs support both parts of that journey, with enough breadth for universal discovery and enough depth for more advanced, product-specific experiences.

Overview

This guide explains how Tuned Global supports content discovery across music and audio products.

It shows how to build discovery experiences using Tuned Global's Metadata APIs, from broad catalogue search through to more precise track retrieval and recommendation-driven continuation.

It should help you understand:

• How users can search for tracks, artists, albums, playlists, stations, and other entities
• How global search, advanced search, and recommendations work together
• How search supports browse, editorial pages, carousels, tags, and recommendation modules
• Which discovery patterns are best suited to different product types
• How discovery supports engagement, retention, and stronger product experiences

Search in Tuned Global is not limited to a single keyword endpoint. The platform supports:

• A broad global search layer for universal catalogue discovery
• A deeper advanced song search layer for more precise track retrieval
• Recommendation flows based on track relationships, tags, and seeded IDs
• Similarity and matching utilities that help extend discovery and improve result quality

How search and recommendation fits into the product

In most products, search is the first step in a wider journey:

Search  -->  Get Results  -->  Playback / Add to Playlist / Get Recommendations Search  -->  Get Results  -->  Playback / Add to Playlist / Get Recommendations

A strong implementation does more than return matches. It:

• Supports different levels of search precision
• Returns enough metadata to power the next UI step
• Helps users move from known-item search into broader discovery
• Connects naturally into playback, artist pages, album pages, playlists, stations, and recommendation modules

Tuned Global's discovery stack is designed to support full discovery journeys across a catalogue, from universal search and track-first retrieval through to browse-led discovery and deeper recommendation flows.

The strongest discovery experiences use these layers together. A practical pattern is:

This creates a discovery stack that supports both known-item search and deeper continuation once the user is engaged.

Global Search

Endpoint: GET /api/v2.4/search Swagger: Metadata API — Search

Global Search is the best starting point when the product needs a universal search bar and the user may be looking for different kinds of content.

This endpoint provides the broadest response, while the other search endpoints are more specific. It searches across the available catalogue and reflects licensing and territory availability, including an available territory node in the response. That makes it well suited to real-world music products where availability matters as much as relevance.

Global Search returns results that can move directly into richer metadata retrieval, playback, browse surfaces, and downstream recommendation flows. It supports broad discovery while still respecting catalogue availability and territories, which makes it suitable for commercial implementations across multiple markets.

Advanced Song Search

Endpoint: GET /api/v2.4/search/songs/advanced Swagger: Metadata API — Advanced Song Search

Advanced Song Search is the deeper track discovery layer. It moves beyond a basic keyword search and supports products that need greater precision, filtering, and richer query behaviour.

This endpoint returns highly structured song-level results with substantial metadata depth, including scoring and ranking fields, identifiers, artist data, label and owner fields, tags, countries, translations, plugin metadata, and release context. That makes it useful not only for retrieval, but also for ranking, filtering, merchandising, and more deliberate discovery workflows.

Advanced Song Search is built for implementation depth, giving teams more room to refine discovery, ranking, filtering, and merchandising.

Recommendations

Search helps the user find something they already want. Recommendations help the product continue the journey once that intent has been expressed.

Tuned Global exposes multiple recommendation patterns rather than a single generic recommendation endpoint. That gives teams flexibility to support different types of continuation across the same catalogue.

Similar artists

Endpoint: GET /api/v2.4/artists/{id}/similar Swagger: Metadata API — Artist

Returns artists similar to a specified artist based on overlapping listening patterns and catalogue attributes. This is the primary similarity endpoint in the Metadata API and supports artist-led continuation when the user's intent is performer-led rather than track-led.

This is especially useful when you want:

• Artist-page continuation
• Fan-led journeys
• Performer exploration
• Artist-based recommendation modules

Similar artists by tags

Endpoint: GET /api/v2.4/artists/{id}/similarbytags Swagger: Metadata API — Artist

Returns artists similar to a specified artist based on tag overlap rather than behavioural data. Useful when you want tag-driven similarity without relying on listening patterns.

Similar stations

Endpoint: GET /api/v2.4/stations/{id}/similar Swagger: Metadata API — Station

Returns stations that share the most tracks by the same artists. Useful for radio-style continuation and "more like this" station modules.

Similar podcasts

Endpoint: GET /api/v2.4/podcasts/{id}/similar Swagger: Metadata API — Podcast

Returns similar podcasts based on matching channel tags.

Playlists by track

Endpoint: GET /api/v2.4/playlists/by-track Swagger: Metadata API — Playlist

Returns playlists that contain a specific track. This is useful for track-led discovery — once a user selects or plays a song, you can surface playlists that feature it, creating a natural bridge from a single track into a broader listening experience.

Tag-driven recommendations

Endpoint: GET /api/v2.4/tags/artists?tag={tag}, GET /api/v2.4/tags/albums?tag={tag}, GET /api/v2.4/playlists/by-tags Swagger: Metadata API — Tag / Playlist

Returns content filtered by tags associated to tracks, albums, or artists. This approach uses the tag system to determine similarity rather than behavioural data.

This is useful when you want:

• Related content after a user views or plays something
• Tag-driven continuation
• Recommendation rows that stay close to the source item's musical profile

TunedIQ — Personalised track recommendations

Reference: Search & Recommendations guide — TunedIQ

TunedIQ is Tuned Global's collaborative-filtering recommendation engine. It learns from real listening behaviour rather than static metadata, making it effective even for catalogues with limited or inconsistent tagging.

Once enabled, TunedIQ provides track-level recommendation capabilities including:

• Seeded recommendations — recommended tracks based on up to 3 specified track IDs
• Tag-driven related tracks — recommendations using tags associated to tracks or derived from albums and artists
• Track similarity — related tracks with rich metadata including identifiers, ISRC, explicit flags, image fields, and linked release data
• Daily Discovery Mixes — blends of familiar favourites and new tracks
• Personalised Trending — trending tracks filtered by the user's taste profile
• Continuous Play — automatically extends listening sessions from a seed track

These are useful when you want:

• A "similar tracks" or "you might also like" module
• Seeded recommendations from a small set of known tracks
• Smarter continuation from a user's recent behaviour
• Recommendation flows that feel more intentional than simple nearest-neighbour similarity

What's available without TunedIQ

The following similarity and discovery endpoints are available in the public Metadata API and do not require TunedIQ:

There is currently no track-level similarity endpoint (e.g. "similar songs") in the public Metadata API. Track-level recommendations require TunedIQ.

Summary

Tuned Global supports multiple recommendation models:

• Similar artists via GET /api/v2.4/artists/{id}/similar (public API)
• Similar artists by tags via GET /api/v2.4/artists/{id}/similarbytags (public API)
• Similar stations via GET /api/v2.4/stations/{id}/similar (public API)
• Similar podcasts via GET /api/v2.4/podcasts/{id}/similar (public API)
• Playlists by track via GET /api/v2.4/playlists/by-track (public API)
• Tag-driven related content via tag discovery endpoints (public API)
• Seeded track recommendations, track similarity, and personalised discovery via TunedIQ (requires activation)

This makes it easier to design recommendation experiences that fit the product.

Additional discovery tools that strengthen the stack

Beyond search and recommendations, Tuned Global also includes supporting discovery tools that help improve result quality, handle catalogue matching, and extend discovery into richer product experiences.

ISRC Matching

Endpoint: GET /api/v2.4/search/search-matching Swagger: Metadata API — Song Matching

This endpoint returns tracks that share the same ISRC. It supports matching by ISRC, song name, artist name, and duration — useful for prioritising a single track per ISRC when multiple tracks exist.

This is useful for:

• Cleaner search results
• De-duplication workflows
• Metadata reconciliation
• Catalogue operations and matching

Display-Ready Similarity Responses

The similar content endpoints (GET /api/v2.4/artists/{id}/similar, GET /api/v2.4/artists/{id}/similarbytags, GET /api/v2.4/stations/{id}/similar, GET /api/v2.4/podcasts/{id}/similar) include not only identifiers, but also artist arrays, release objects, translations, image fields, and other display-ready data. Recommendation modules can therefore be wired directly into product experiences without needing a separate minimal similarity layer.

Supporting Discovery Surfaces

Search and recommendations become even more powerful when combined with supporting discovery surfaces such as:

• Content pages
• Carousels
• Tags (via GET /api/v2.4/search/tags and GET /api/v2.4/tags/{type}?tag={tag})
• Editorial modules
• Browse-led collections

These surfaces help turn discovery into a broader product experience rather than a single search box.

Catalogue filtering and availability

Discovery quality depends on more than relevance alone. Tuned Global's search and recommendation stack can also reflect catalogue availability, licensing, territories, and richer metadata such as tags, genres, language, owner, label, and release context.

This makes it possible to build discovery experiences that are not only relevant, but also commercially usable in the markets and contexts where the product operates.

Getting started

A practical way to start implementing search and recommendation capabilities is:

When to use each endpoint

Why this is powerful

Tuned Global's discovery capability is strong because it is not built around only one search pattern. It combines:

• A broad universal search layer (GET /api/v2.4/search)
• A deeper advanced track-search layer (GET /api/v2.4/search/songs/advanced)
• Entity-level similarity in the public API — artists (GET /api/v2.4/artists/{id}/similar, /similarbytags), stations (GET /api/v2.4/stations/{id}/similar), podcasts (GET /api/v2.4/podcasts/{id}/similar)
• Track-to-playlist discovery (GET /api/v2.4/playlists/by-track)
• Tag-driven browse and recommendation surfaces
• ISRC and metadata matching (GET /api/v2.4/search/search-matching)
• Track-level personalised recommendations via TunedIQ (separately licensed)
• Rich metadata responses that are already useful for downstream UI and playback flows

That gives technical teams more freedom to design discovery experiences around product intent, not just around whichever endpoint exists.

Swagger References:

• Metadata API (v2.4) — Search, discovery, tags, trending, similar content
• Services API (v3) — Play history, personalised playlists, authenticated similar artists