Research report · July 2, 2026 · Cloudflare-compatible Inworld voices

Mudo voice catalog mapping for Inworld TTS 1.5 Mini

This report reads data/voice-catalog-cache.json. With an Inworld API key, that cache can contain descriptions, tags, gender, age group, categories, and free preview audio. Without the key, it falls back to the 95 Cloudflare-compatible voice IDs.

Cloudflare voices -

Metadata matched -

Free previews cached -

Cache status -

How This Should Work In Mudo

Backend job calls Inworld GET /voices/v1/voices with INWORLD_API_KEY and filters for source = "SYSTEM".
Backend joins Inworld voiceId values to the Cloudflare voice_id enum for inworld/tts-1.5-mini.
Backend caches the joined catalog. This report cache lives in reports/inworld-voice-catalog-2026-07-02/data/voice-catalog-cache.json; product should use KV/R2/D1 with a TTL or manual refresh.
Voice previews come from Inworld GET /tts/v1/voice:preview. Inworld documents this as free and unbilled. The browser should call our backend cache/proxy, not Inworld directly.
Cloudflare remains the synthesis provider: selected voiceId becomes the Cloudflare voice_id parameter for inworld/tts-1.5-mini.

Voice Browser

Search Gender Age group Preview

Sources

Cloudflare Inworld TTS 1.5 Mini docs: Cloudflare-compatible voice enum and synthesis parameters.
Inworld List Voices API: voice metadata fields and filtering.
Inworld Get Voice Preview API: free/unbilled preview endpoint returning MP3 bytes.
Inworld Voice Tags: tags for gender, age, style, and quality.