K

Text-to-Speech Software

KugelAudio Text-to-Speech Review

kugelaudio.com · checked 11 Oct 2026

Text-to-speech platform for real-time apps, voice agents, and content tools.

How it plugs in

YesPublic API
NoOpen source
YesFree tier

Signal on Text-to-Speech Software7.5#61 of 223

KugelAudio delivers multilingual speech synthesis for live voice agents, interactive software, and content workflows. It offers web, API, and self-hosted options, with voice cloning, audio streaming, and exports in common formats. Speech-to-text is also documented. A free testing plan is available, with paid models billed by generated audio.

KugelAudio home page

Review

In a line

KugelAudio provides speech synthesis for applications that operate in real time. It can generate speech for voice agents, interactive software, and content creation tools.

What it does

Its multilingual model supports 39 languages and can clone voices. WebSocket streaming sends audio pieces as text arrives, which suits workflows that receive text token by token.

The documentation also covers speech-to-text using the luchs-1 model. Generated audio can be exported as WAV, PCM, MP3, µ-law, or a-law files.

How it connects

KugelAudio runs through a web interface, API, or self-hosted deployment. Developers can call REST or WebSocket endpoints, with quickstarts covering Python, TypeScript/JavaScript, and Java.

For voice-agent integrations, the documentation points to LiveKit, Pipecat, Vapi, and Cognigy. Enterprise customers can run KugelAudio on their own infrastructure while retaining the SDK-facing interface.

Hosted speech processing runs in configured EU regions. EU and EEA customers can use EU Sovereign Cloud. Traffic uses TLS 1.2 or higher.

Good fit

KugelAudio suits teams building voice agents, interactive applications, or content tools that need speech as text is produced. Voice cloning and multilingual output help when a project needs a consistent voice across many languages.

Its API choices and language quickstarts fit developers, while the web option provides a way to use the service without hosting it yourself.

Watch for

The Trust Center describes GDPR compliance as self-declared, not certification from the EU or another third party. Text, audio, prompts, and transcripts are handled transiently and removed after the request; voice-reference retention is governed separately.

Enterprise deployment terms, availability, and support responsibilities are agreed for each installation. Check those details if you need on-premise hosting or custom concurrency.

Pricing

A free plan provides testing access, and the site says you can start without a credit card. Paid access starts at €0.04.

The kugel-2-turbo plan is priced at €0.04 one-time, calculated per generated second while displayed per audio minute, before VAT. Its listed VAT-inclusive amount is €0.04165. kugel-3 costs €0.07 one-time on the same basis, or €0.0833 including 19% VAT.

Both paid models list EU hosting, voice cloning, speed control, break tokens, word-level timestamps, dictionaries, and EU Sovereign Cloud for EU/EEA customers. kugel-3 also lists IPA notations. Enterprise pricing is on request, with custom rates and lower rates for committed usage.

Verdict

Choose KugelAudio if your project needs real-time multilingual speech, voice cloning, or streaming audio through an API. The free testing access and several connection methods make it possible to evaluate before paying.

Look more closely at retention rules, self-declared GDPR status, and deployment responsibilities if your data or hosting requirements are strict. Enterprise buyers should request details on support, concurrency, and custom deployment terms.

Our take

KugelAudio is a reasonable choice for teams creating voice agents or interactive software that needs speech generated while text is still arriving. It covers 39 languages, voice cloning, common audio exports, and both REST and WebSocket access. The free testing plan lowers the barrier to evaluation. Buyers with strict compliance needs should review its self-declared GDPR position and the separate rules for stored voice references. On-premise customers also need to settle availability and support terms directly for each deployment.

  • Supports 39 languages
  • Voice cloning and real-time audio streaming
  • REST and WebSocket APIs
  • Free testing plan without a credit card
  • EU hosting and Sovereign Cloud options
  • GDPR compliance is self-declared
  • Voice-reference retention is governed separately
  • Enterprise deployment terms vary by agreement
  • Paid prices are shown per audio minute and billed by second
  • Pick it ifReal-time voice-agent and interactive software projects; Content tools needing multilingual speech; Teams seeking API or self-hosted deployment
  • Skip it ifYou require third-party GDPR certification; You need fixed on-premise support terms before contracting; Your project has no need for speech generated in real time

Pricing

Free testing access is available. Enterprise pricing is on request.

  • Free tierTesting accessFree
  • kugel-2-turboPer generated second; price shown per audio minute; excl. VAT · €0.04165 incl. 19% VAT · ~40–50 ms TTFA · Guaranteed EU hosting · Voice cloning · Speed control · Break tokens · Word-level timestamps · Dictionaries · EU Sovereign Cloud for EU/EEA customers€0.04 one-time
  • kugel-3Per generated second; price shown per audio minute; excl. VAT · €0.0833 incl. 19% VAT · ~60–80 ms TTFA · Guaranteed EU hosting · Voice cloning · Speed control · Break tokens · Word-level timestamps · Dictionaries · IPA notations · EU Sovereign Cloud for EU/EEA customers€0.07 one-time
  • EnterpriseCustom rates; lower rates with committed usage · On-premise or dedicated deployment · Higher custom concurrency limits · 24/7 support · Dedicated account managerOn request

As kugelaudio.com lists them, checked 20 Sept 2026.

Questions

Is there a no-cost KugelAudio tier?

Yes. The Free tier provides testing access, and the homepage says you can start without a credit card.

What language coverage does KugelAudio provide?

Its multilingual text-to-speech model supports 39 languages.

Which interfaces and SDK quickstarts are available?

REST and WebSocket endpoints are available, with quickstarts for Python, TypeScript/JavaScript, and Java.

Can KugelAudio clone voices?

Yes. Voice cloning is listed for the service and its paid model plans.

Where is KugelAudio hosted?

Hosted speech processing runs in configured EU regions. EU and EEA customers can use EU Sovereign Cloud.

How are submitted text and audio handled?

Text, audio, prompts, and transcripts are processed transiently and deleted after the request. Voice references follow separate storage and deletion rules.

Can businesses run KugelAudio themselves?

Yes. Enterprise customers can deploy it on their own infrastructure using the same software-development interface; terms are agreed for each deployment.