TTS engine multipliers explained: 1.0× to 3.2× — Audra

Audra bills one prepaid base — $25 per million characters at checkout — and each engine applies a fixed multiplier on top. There are four multipliers, they never scale with volume, and this post prices the same script on each so the choice is a lookup, not a surprise invoice.

The multiplier table

• Essential (audra-essential) — 1.0× — $25.00/1M — the batch narration default, 72 voices across 10 languages.
• Plus (audra-plus) — 1.15× — $28.75/1M — extended multilingual CPU presets.
• Pro (audra-pro) — 2.0× — $50.00/1M — GPU MOSS narration for module intros and ad-style reads.
• Conversations (audra-conversations) — 3.2× — $80.00/1M — multi-speaker shows, 2–5 speakers in one request.

Fast mode adds a further 1.25× for queue priority on top of the engine multiplier, and it is opt-in per request.

The same 500k-character block, priced four ways

Choosing by workload, not by ceiling

• Narration that people listen to for comprehension — docs readers, course bodies, podcast narration — Essential is the default and the cheapest.
• Marketing-adjacent audio where a flat read hurts conversion — sample Pro on the same script before committing the catalog.
• Dialogue-shaped content — roundtables, co-hosted segments — Conversations replaces stitching files in ffmpeg, which is where its 3.2× earns back.

Engine capability detail lives at /blog/audra-four-engines-guide; the full rate card at /pricing#engines.

Next steps

• Start here: /blog/audra-four-engines-guide
• Then: /calc
• Reference: /docs/engine
• More Audra guides: /blog