Tool overview
LMNT is listed under Voice AI tools.
What is LMNT?
LMNT provides low-latency text-to-speech and voice-cloning APIs for conversational agents, games, media, and accessibility experiences. Blizzard 2.0 supports streaming, speech sessions, word timestamps, accent control, and 31 languages with native code-switching.
Best for
Developers building multilingual realtime voice and speech experiences
Who is it for?
Decision note
Accepted for Preview after independent V411 official-source research. Apply only after failures = 0, warnings = 0, unmapped = 0, and all proposed changes are reviewed.
Key features
Streaming text-to-speech with 150–200 ms latency
Voice cloning from a short reference recording
Thirty-one languages with native code-switching
Speech sessions, accents, word timestamps, and official SDKs
Use cases
Give realtime agents natural voices
Create localized voice experiences
Generate narration and game dialogue
Build branded custom voice applications
Pros
- Public free and paid API tiers
- Supports Arabic and thirty other languages
- Built for low-latency streaming
Cons
- Character allowances and overage rates vary by plan
- Voice cloning still requires suitable reference audio and consent
Limitations
Output quality depends on prompts, language, and reference recordings
Commercial use requires an eligible paid or enterprise plan
Pricing details
Billing options
Pricing note
Free includes 15,000 characters and unlimited voice clones. Indie costs $10 per month for 200,000 characters, Pro $49 for 1.25 million, and Premium $199 for 5.7 million; overage rates decline by tier. Enterprise and startup-grant pricing are available separately.
Supported languages
- 31 languages including Arabic, English, Spanish, French, German, Hindi, Japanese, and Chinese
Integrations
LiveKit
Pipecat
Vapi
Vercel AI SDK
Please log in to join the discussion.