Toku Launches Makimoto: Open-Source AI for Asia-Pacific Data Residency

Toku, a Singapore Exchange Catalist-listed AI platform, has announced Makimoto, an open-source conversational AI initiative designed specifically for Asia-Pacific data residency requirements. The first release, Makimoto Kawa, launches as a managed transcription API service on 1 July 2026, hosted exclusively in Singapore.

The announcement addresses a critical engineering challenge facing enterprises across the region: where customer audio and transcripts are processed, not just stored. The official press release details how data-processing regulations across Asia-Pacific are fragmenting, with Indonesia's Personal Data Protection Law, Vietnam's Personal Data Protection Law, and Singapore's updated PDPA all raising stakes on jurisdiction-specific processing.

For teams building voice and chat automation, the primary engineering question has shifted. It is no longer which model delivers the highest accuracy. It is which deployment satisfies each regulator while scaling across markets. Makimoto is designed to answer that question on day one.

Thomas Laboulle, Founder and CEO at Toku, frames the initiative as a response to technology obsolescence. "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's composable architecture allows every component to be swapped as the field evolves (a problem that has plagued users for years, frankly).

The orchestration layer itself is what Toku open-sources under the MIT licence. This is the glue code, configuration, and APIs that allow every component in the pipeline to be swapped without rewriting the surrounding system. Kawa will be offered through two APIs at launch: one for real-time transcription for live voice agent experiences and live captioning, and one for post-conversation transcription suited for recorded calls and batch analytics.

Under the surface, both APIs run the same five-stage modular pipeline. Audio resampling comes first. Then noise filtering and audio enhancement. Voice activity detection with speaker diarisation follows. Speech-to-text inference comes next. Post-processing handles normalisation, temporal ordering, speaker labelling, and structured output. Operators can swap or tune any individual stage to optimise for a given language, domain, or deployment environment.

At launch, the orchestration layer opens as MIT-licensed source code. Internal components remain managed by Toku. These include voice activity detection, diarisation, noise handling, the speech-to-text model, and post-processing. Over subsequent releases, Toku will progressively open additional layers of the pipeline. The goal is full transparency and community modifiability of every component.

This phased approach lets the Makimoto community review, contribute to, and eventually replace each part of the stack. Meanwhile, Toku maintains production stability for early-access users. The waiting list for early access opens at makimoto.ai.

The official Makimoto website confirms the progressive open-source timeline. Internal pipeline components will open progressively through 2026 and 2027. There are no source-available carve-outs. There are no Business Source License surprises. Toku commits to keeping Makimoto's open-source components permissively licensed.

Makimoto is operated through Makimoto Technology Pte Ltd, a wholly-owned subsidiary of the Toku Group. This structure preserves community trust and transparent governance. It maintains a clear distinction between enterprise and community editions of Kawa. The community edition evolves under open governance and external contribution. The enterprise edition stays focused on reliability, security, and commercial support for regulated-sector customers.

Over time, additional components of Toku's technology stack will migrate into Makimoto. This evolves the group's architecture toward an open-core approach. It reinforces the composability thesis at the heart of this initiative.

Alongside the early-access programme, Toku is opening ten Singapore-based positions on the Makimoto team. The roles are open to recent graduates of Singapore's universities and polytechnics. They span engineering, developer relations, and community work. The hiring programme reflects Toku's conviction that developing future-ready talent matters as much as the technology itself.

The physical reality of using Kawa differs from typical AI services. Developers interact with a pipeline where every stage is replaceable. They can tune for their language. They can tune for their domain. They can tune for their latency budget. Composability is the architecture, not a feature.

Customer audio and transcripts are processed in country, not just stored there. Kawa's first release runs exclusively in Singapore. Country-specific deployments follow. A self-hostable container release follows. This matters for jurisdictions where data residency is non-negotiable.

Whether enterprises actually adopt this model remains the real question. The regulatory landscape is complex. The technical requirements are demanding. The open-source community will need to engage meaningfully for this to work as intended.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn