Skip to main content
Rime exposes a broad portfolio of production-ready voices across our models. Each voice is designed for real-time conversational performance, with distinct tonal identity, emotional range, and demographic diversity. As of 4 February 2026, Rime supports the following languages:
  • Arabic lang=ara or lang=ar (Arcana only)
  • English lang=eng or lang=en
  • French lang=fra or lang=fr
  • German lang=ger or lang=de
  • Hebrew lang=heb or lang=he (Arcana only)
  • Hindi lang=hin or lang=hi (Arcana only)
  • Japanese lang=jpn or lang=ja (Arcana only)
  • Portuguese lang=por or lang=pt (Arcana only)
  • Spanish lang=spa or lang=es

​
Arcana v3 Voices (Flagship)

Arcana v3 is Rime’s most expressive and human-realistic text-to-speech model. The voices in Arcana are designed to sound natural in live conversation — capturing pacing, rhythm, breath, subtle emotional shifts, and multilingual fluency without sacrificing responsiveness. Arcana v3 supports native multilingual speech and code-switching.

​
The 94 Arcana Flagship Voices

Arcana v3 includes 94 flagship voices spanning a wide tonal and demographic spectrum. Rather than optimizing for a single “neutral assistant” sound, Arcana voices are intentionally distinct and characterful. At a high level, the range includes:
  • Age diversity — young adult through older adult voices
  • Regional variety — American regional accents (Southern, Midwestern, West Coast, East Coast), as well as international English and native speakers across supported languages
  • Cultural representation — African American, Latina/o, Asian American, and other globally representative identities
  • Gender spectrum — balanced representation across male and female voices, including stylistic variation within each
  • Energy profiles — from calm and grounded to upbeat, playful, warm, authoritative, or highly conversational
  • Use-case fit — voices well-suited for IVR, customer support, healthcare, fintech, education, media narration, and expressive AI companions
Arcana voices are built to maintain voice identity even when switching languages mid-utterance, making them especially powerful for global conversational applications.

​
Mist v2 Voices

Mist v2 is optimized for speed, precision, and control. While Arcana emphasizes expressive realism, Mist focuses on clarity, pronunciation accuracy, and high-throughput production environments. Mist voices are:
  • Highly intelligible and consistent
  • Well-suited for structured IVR systems and high-volume applications
  • Optimized for fast synthesis and predictable delivery
  • Designed to handle complex proper nouns, brand names, and domain-specific terminology cleanly
Mist v2 provides a curated set of dependable, production-stable voices that prioritize control and scalability.

​
Accessing Voices

Detailed voice metadata is available: Use the speaker parameter in the TTS API to select a voice.