Toku, a Singapore Exchange (SGX) Catalist-listed AI-powered customer experience platform, has launched Makimoto, an open-source conversational AI initiative designed to keep customer data within national borders across Asia-Pacific. The first release, Makimoto Kawa, will be available from 1 July 2026 as a managed transcription API hosted in Singapore.
The launch is a direct response to tightening and diverging data-processing regulations across APAC. Indonesia’s Personal Data Protection Law, Vietnam’s Personal Data Protection Law, and Singapore’s Personal Data Protection Act have raised the stakes for enterprises processing audio, transcripts, and customer conversations across multiple markets.
“In an era where AI capabilities evolve on a weekly cycle, the biggest risk of any technology decision is creating tomorrow’s legacy today. Makimoto is our response: a composable architecture where every component can be swapped as the field evolves.” — Thomas Laboulle, Founder and CEO, Toku
What Makimoto Kawa does
Makimoto Kawa is an MIT-licensed orchestration framework that runs a five-stage modular pipeline — audio resampling, noise filtering and enhancement, voice activity detection with speaker diarisation, speech-to-text inference, and post-processing — as a production-grade transcription service. The orchestration layer, which connects and configures the underlying components, is open-sourced; individual pipeline components remain managed by Toku at launch, with additional layers to be opened progressively.
At launch, Kawa offers two APIs: one for real-time transcription suited to live voice agents and in-conversation analytics, and one for post-conversation transcription covering recorded calls, voicemail, and batch processing. Operators can swap or tune any stage of the pipeline to accommodate different languages, domains, or deployment environments.
“With Kawa, customer data is processed in country, not just stored there. Every stage of the pipeline runs within a single jurisdiction, so our customers always know exactly where their data is going.” — Thomas Laboulle, Founder and CEO, Toku
Data residency by design
Kawa’s initial release runs exclusively in Singapore, supporting in-country processing for customers subject to the PDPA and regulated-sector guidelines from the Monetary Authority of Singapore. Toku says Makimoto is structured to address markets such as Indonesia and Vietnam, which enforce strict in-country data-processing mandates for real-time communications and transcription services.
A containerised, self-hostable version will follow the managed API release, enabling deployment in a customer’s own cloud, data centre, or jurisdiction. Makimoto is operated through Makimoto Technology Pte Ltd, a wholly-owned Toku subsidiary, to maintain separate governance between community and enterprise editions. Toku has committed to keeping open-source components under permissive licensing, with no plans to move to source-available or Business Source License terms.
Ecosystem interoperability and talent
Toku says Makimoto is designed to interoperate with regional language models including SEA-LION from AI Singapore and MERaLiON from A*STAR, which developers can integrate as reasoning or dialogue layers in applications requiring accurate transcription of regional speech.
Alongside the launch, Toku is opening ten Singapore-based graduate positions spanning engineering, developer relations, and community work. Applications and the early-access waiting list are open at makimoto.ai.
