Jutusta

Voice · Transcription · Translation

Everything for speech,
in one place

Generate natural speech. Transcribe live or from files. Translate speech in real time. All in 45 languages.

Start transcribing Try voice synthesis

45 languages · Developer API · Self-hostable

🇹🇭 th
🇹🇷 tr
🇺🇦 uk
🇵🇰 ur
🇻🇳 vi
🇭🇰 yue
🇨🇳 zh
🇪🇪 et
🇸🇦 ar
🇦🇿 az
🇧🇬 bg
🇧🇦 bs
🇪🇸 ca
🇨🇿 cs
🇩🇰 da
01 / 45 · Eesti Estonian

Measured Accuracy

More accurate in Estonian than the world's biggest

On a public Estonian speech-recognition benchmark, jutusta.ee turns speech into text with a lower error rate than the systems from ElevenLabs, Google, Microsoft, and OpenAI.

Word error rate lower is better
jutusta.ee 6.2%
ElevenLabsScribe v2 8.8%
GoogleGemini 3.5 Flash 9.0%
MicrosoftAzure Speech 14.2%
GoogleGemini 3.1 Pro 14.4%
OpenAIWhisper large-v3-turbo 23.9%
OpenAIGPT-4o Transcribe 27.4%

Measured in July 2026; all systems scored with the same methodology. Benchmark: github.com/alumae/k6net6lke-benchmark

Streaming Transcription

Words appear as you speak

Low-latency speech recognition in 39 languages. Punctuation and capitalisation are added automatically.

Try it free →
Recording
ET Eesti

Voice Synthesis

Hear the difference

Natural built-in voices in 45 languages, generated from any text.

Heli

Estonian · female voice

“Welcome to Jutusta — hear how natural this voice can sound.”

Try voice synthesis →

Speech Translation

Translate in real time

Speak in one language, read or listen in another — in real time. 39 source languages, 45 target languages.

ET Eesti Real-time
EN English
Try translation free →

Everything you need in one place

01

Voice Synthesis

Generate natural-sounding speech from text. Tune gender, pitch, age, and style — or clone a voice from a short sample.

02

Streaming Transcription

Speech-to-text as you talk. Low latency, punctuation, and strong support for Estonian.

03

Offline Transcription

Upload audio or video for accurate batch transcription. Works for longer recordings too.

04

Speech Translation

Translate speech in real time or from files, across 45 languages.

Developer API

Plug speech AI into your product

An ElevenLabs-compatible HTTP API for TTS, plus a streaming and batch ASR API. EU-hosted, simple API-key auth, and a live OpenAPI reference.

Open the API docs →
$ POST /api/v1/text-to-speech/:voice_id 
curl https://jutusta.ee/api/v1/text-to-speech/<voice_id> \
  -H "xi-api-key: $JUTUSTA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"text": "Tere maailm", "model_id": "jutusta_v1"}' \
  --output speech.mp3

Available now

Turn Microsoft Teams recordings into meeting minutes

Connect your Microsoft work or school account and have Jutusta automatically transcribe and summarize finalized Teams recordings. Usage is charged from your Jutusta credits, and you choose how meeting data is stored during setup.

Sign in to connect Teams

After sign-in, you will land on the Teams setup form.

Available now

Turn Google Meet recordings into meeting minutes

Connect your Google Workspace account and have Jutusta automatically transcribe and summarize finished Meet recordings. Usage is charged from your Jutusta credits, and you choose how meeting data is stored during setup.

Sign in to connect Meet

After sign-in, you will land on the Meet setup form.

Roadmap

Coming soon

The next products and deployment modes for Jutusta.

  1. Q3 · 2026 Coming soon

    Hosted in Riigipilv

    A dedicated environment in Riigipilv (the Estonian government cloud) for public bodies and organisations with strict data-protection requirements and data-residency constraints.

  2. Q3 · 2026 Coming soon

    On-premise install

    Deploy Jutusta on your own servers or in an internet-isolated environment. Your audio and data never leave the building. Jutusta.ee's own back office already runs self-hosted.

Ready to speak?

Free to try in the browser. Self-hosted deployments available — talk to us.

Start free Generate speech