About Jutusta

Speech technology with research roots and production habits.

Jutusta is developed by Next Interaction OÜ in Tallinn. The product brings speech synthesis, speech recognition and translation into one practical environment, with special attention to Estonian and other smaller languages.

Behind the product is Aivo Olev, a TalTech NLP researcher and PhD student whose work connects academic speech AI with systems people can actually use.

Mission

Jutusta exists to bring research-based Estonian language technology to companies, public institutions and private users. The ambition is practical access without lowering the bar: speech recognition, speech synthesis and translation should each be developed with a direct line to frontier research, not only wrapped around generic models.

Company

Next Interaction OÜ develops Jutusta and builds speech and language technology for organisations that need more than a standard web app: integrations, private deployments and clear control over where data is processed.

Local hosting

As a matter of principle, we do not use cloud-based large language models for user content, and we refuse to send user data to the United States or China. Speech and language technology must be useful without making the user give up control over their recordings and text.

Aivo Olev

Aivo started working in software in 2008 as a systems analyst, later adding a BA in Psychology and a cum laude Computer Science degree from TalTech. That path still shapes the work: product decisions are treated as both technical and human problems.

At TalTech, his recent focus has been speech recognition, audio reasoning, language identification and agentic AI systems. The goal is not to make AI sound impressive in a demo, but to make it reliable enough for real recordings, real accents and real workflows.

Selected 2026 references

A few public references show the research and engineering work behind Jutusta.

arXiv, 18 Mar 2026 Multi-Source Evidence Fusion for Audio Question Answering Aivo Olev and Tanel Alumäe describe TalTech’s first-place Agent Track system for the Interspeech 2026 Audio Reasoning Challenge.Interspeech 2026 leaderboard TalTech ranked first in the Agent Track The official leaderboard lists TalTech first in the Agent Track, ahead of teams from major research labs and technology companies.TalTech, 2026 Researcher and builder profile TalTech describes Aivo as a PhD student, NLP researcher and experienced IT specialist, connected to tekstiks.ee and Jutukuva.Public profile Odyssey 2026 TidyLang and recent work Aivo’s public profile references the 2026 TidyLang language identification challenge and recent speech/audio competition work.

Open-source and public projects

tekstiks.ee Open-source Estonian speech transcription platform created at TalTech, with an editable transcript workflow and public references to the underlying ASR components.Jutukuva Free real-time Estonian speech recognition app for private local use, live correction and sharing. Built with EKI and TalTech NLP.GitHub: aivo0 Additional public experiments and engineering work.

For teams and integrations

If you need speech tools in a product, a private deployment, or a workflow around Estonian speech data, we can help with integrations and deployments outside the public web app.