Roadmap

Add AI powered routes that are mostly just for fun, plus a few genuinely useful ones for radio owners, the general public and so on. Grounded on real API data, cached for 30 days, and rate-limited hard so they stay cheap.

Hardening how the AI assistant handles a busy or unresponsive model: health-aware failover, faster timeouts and automatic retries, so a slow backend never leaves you waiting.

Move beyond IP-only rate limiting. Keys unlock per-key quotas, usage dashboards, and the ability to revoke or rotate credentials.

Request exactly the fields you want across entities in one query. Cuts payload size and round-trips, building on the existing /batch foundation.

Emit Deprecation and Sunset headers on ageing endpoints, with a documented support window so you know when something will actually go away.

Thin typed clients for JS/TS and Python, auto-generated from the OpenAPI spec so they never drift from the live API.

Widen /track/audio-analysis coverage across the catalogue: tempo, key, energy, danceability and more, enabling mood and vibe-based discovery.

Given a seed set, stream an endless recommended queue. Like /artist/radio but multi-seed, paginated and designed for building listening apps.

Accept a list of ISRCs or streaming-service URLs in one call and return canonical records for all of them. A natural extension of /resolve + /batch.

Pin any path to a frozen API version. Old versions stay exactly as they were, so nothing we ship later can move the ground under you.

Every error across the whole API now returns the same shape, with a stable code you can branch on and a human-readable message.

RateLimit-Limit, RateLimit-Remaining and RateLimit-Reset following the IETF draft, alongside the old X-RateLimit-* headers for backwards compatibility.

The assistant can read the full version history now. Ask it what changed since any version or what might break your integration, and it pulls the real notes.