Uses
Create voice agents through APIs. Generate speech from text. Transcribe audio in real time. Build speech-to-speech applications with a dedicated model.
Tool profile · Publisher documentation
Smallest AI
Smallest AI is presented by its supplier as a realtime voice AI suite built for enterprise scale, offering APIs to create voice agents, generate speech and transcribe audio in real time. LE LABO IA has not tested the product, and this profile relies on the public documentation of its supplier. Documented features include text to speech, real-time audio transcription, voice agents, a speech-to-speech model and voice cloning. The main intended use is to build voice agents, generate speech and transcribe audio in real time through APIs. Before committing, you should check unconfirmed prices, quotas, territorial availability and usage terms with the supplier, and confirm the pricing on the day you decide to use it.
Read this page in:Français
Tool profile · Publisher documentation
What to check before choosing
The product points below come from publisher documentation. No common hands-on test is claimed. Documentary profile, sources checked October 7, 2026. LE LABO IA has not tested this product.
Offer and pricing
Pricing is documented as pay as you go. It includes 10 USD of free credits to start, then per-call prices from 0.09 to 0.21 USD per minute depending on models, and as low as 0.05 USD per minute. Smallest hosting is 0.01 USD per minute. Speech to text is about 0.009 USD per minute, and text to speech about 0.09 USD per minute. A phone number costs 10 USD per number per month. Knowledge base pricing is 3 USD per GB per user and 2 USD per 1,000 queries. Personal data deletion is 0.01 USD per minute. Text messages cost 0.005 USD each. Optional zero data retention for HIPAA is 1,000 USD per month. An Enterprise plan is available on quote. These prices are not confirmed; check them with the supplier.
Limits to check
The product description is based on public supplier documentation and no hands-on test was performed. Quotas, territorial availability and usage terms are not confirmed and should be checked with the supplier. The pricing page was read on 2026-10-07, but prices should be re-checked on the day you decide to use the product. Model prices for LLM and speech to speech are indicated as billed at cost, which is a pricing caveat to verify with the supplier.
What to check yourself
Before committing, check on the supplier site or during a trial whether the voice agent, speech generation and transcription APIs support your language pair, latency target and audio format. Confirm the current per-minute prices, free credit conditions, phone number costs and knowledge base limits directly with the supplier. Ask how voice cloning consent, data deletion and HIPAA retention options work in practice for your use case. Verify that the Enterprise quote covers your expected call volume and any territory restrictions.
- Prices, quotas, territorial availability and usage terms are not confirmed and should be checked with the supplier.
- The supplier indicates that LLM and speech to speech model amounts are billed at cost, but this is not confirmed independently.