On-premise or hosted
Speech-To-Text models

State of the art WER

diagram

Available languages

GBEnglish
NLDutch
FRFrench
PTPortuguese
ESSpanish
DEGerman
ITItalian
plusReady soon:
ILHebrew

Affordable Pricing

BETA

PAY-AS-YOU-GOHOSTED API

$0.2 / hour
$0.0033 / min
TRY IT FOR FREE

*NO CREDIT CARD REQUIRED

  • check$10 credit, on us. Then pay-as-you-go. No expiration.
  • checkAll supported languages, streaming and offline, same price.
  • checkCloud Hosted by Banafo.
  • checkWebsockets API.
  • checkup to 10 concurrent requests (more available at no extra cost).
  • checkLow latency.
  • check10 access tokens.
BETA

PRIVATE ON-PREMISEMODELS

starting at $500 / month
  • checkPrepaid credit for the month (minimum commitment).
  • checkLowest cost per hour.
  • checkSelf-Hosted on your premise, on your private or public cloud. In docker or on your favourite (recent) Linux distribution (Windows / macOS coming soon).
  • checkWebsockets API.
  • checkHighest Privacy / Security.
  • checkLowest possible latency.
  • checkCPU based, no GPU required.
  • checkup to ~10 concurrent channels per CPU core (without loss of quality).
  • checkStreaming or offline.
BETA

ENTERPRISESOLUTIONS

  • checkFor businesses with tailored needs. Large volumes, custom models, languages, or integration.
  • checkUnique support needs to special pricing for charities / non-profit organizations.
  • checkHosted by choice - Banafo hosted, on your premise or your private cloud.
  • checkCustom models (Different languages / Dialects).
  • checkOn-device (WebAssembly, iOS, Android).
  • checkHigher volumes and concurrency.
  • checkIntegrations with third-party apps and services.

Model Features

verifiedUltimate Privacy. Your hardware. Data never leaves your servers.
verifiedNo GPU needed
verified50x cheaper than Google ASR (starting at 0.01 $/hour)
verifiedState-of-the-art WER (word error rate)
voice_selection

Models for Streaming audio

Great for
Low latency
Live captions
Accessibility
Voice command services
Chatbots
Virtual assistants
View Plans
x10  real-time on a single CPU core
voice_chat

Models for pre-recorded audio

Great for
Meeting transcripts - online and offline
Visual voicemail
Voice memo transcripts
Content creation
Productivity and analytics
Movie subtitles
View Plans
x20  real-time on a single CPU core

Optimized for calls (8kHz or 16kHz)

Models optimized for Call Centers with highest accuracy, for processing recorded speech.
verifiedWebsockets API
verifiedPunctuation
verifiedWord timestamps
verifiedCapitalization
verifiedNo hallucinations

Frequently Asked Questions

Try the models. No login or credit card required.
Demo