Explore how teams are building with MERaLiON

Real products built on MERaLiON across healthcare, finance, public services, and enterprise.

Ajentik

Caregiver assessments and care coordination in Elderwise.ai, supporting families navigating home care through accessible interactions and AI-driven workflows for community and public healthcare.

Axiom

AI Voice Agent platform built on voice network that automates diverse communication tasks such as call centre operations, elderly outreach, and scam call detection, with local language understanding and sentiment analysis.

Broby

Speech-to-text transcription for veterinary consultations involving Singlish and code-switching.

Cloudsquire

Mixed-language speech transcription (Singlish, English, Mandarin blends) of voice-based engagements from users across Singapore, on their very own engagement AI platform, HearBack.

Dez Labs

Dez, an AI-powered preventative health Intelligence platform and application for families managing chronic conditions (NCDs) at home.

iNextLabs

Multilingual ASR across English, Tamil, Mandarin, Malay, and Singlish for real-time voice AI workflows, with low-latency transcription for Southeast Asian accents and conversational speech.

Fatfish

MERaLiON integrated to enable advanced human-robot interaction across real-world deployment scenarios, including use cases at the Science Centre Singapore.

HTX

MERaLiON's Singaporean-accented English capability made it well-suited for HTX's operational use cases, where context and dialect accuracy matter considerably.

NCS

Voice transcription layer in NCSgpt enabling hands-free, multilingual interaction across Southeast Asian languages and accents, letting employees submit prompts via speech for improved accessibility on mobile.

Salmon Thrust & SFX

AI avatar conversational training for bankers, with code-switching and sentiment analysis enabled by MERaLiON for more realistic practice scenarios.

ST Engineering

Decision Support Systems for mission-critical operations, using MERaLiON to transcribe and analyse multilingual audio from operational environments and surface key information faster.

What teams are saying

Practitioners who have integrated MERaLiON into real products, in their own words.

Integrating MERaLiON-CTC-630M into NCSgpt brought voice transcription capabilities that actually understand our multilingual Southeast Asian workforce. The model's ability to handle code-switching and regional accents eliminated the frustration with generic voice recognition tools. We would recommend MERaLiON to any enterprise operating in Southeast Asia that needs voice AI to work seamlessly across English, Mandarin, Malay, and mixed-language conversations — particularly for internal productivity tools where cultural and linguistic nuance matters.

Code-switching is the hardest part of speech recognition in Singapore, and most models we tried struggled with it. MERaLiON was the first one that handled mixed-language speech in a way that felt usable in a real clinical setting. It is a meaningful step forward for anyone building voice-first products for the local market.

For what we're building, having an LLM that can code-switch from Singlish to the primary languages in Singapore has made Dez more accessible. MERaLiON was also more responsive and in alignment with our guardrails, which was a bonus. Working with the MERaLiON team has been effortless.

MERaLiON fills an important gap in the speech-to-text space, particularly for Singapore's unique linguistic environment. The API is straightforward to integrate and performs reliably even with mixed-language inputs that typical STT engines struggle with. We would recommend it to teams working with local speech data or multilingual conversational inputs.

Our experience with MERaLiON-2-10B has been very positive, especially for multilingual ASR use cases involving Southeast Asian languages and mixed accents. The model delivered strong transcription quality with lower latency compared to other transcription models we evaluated in this region, which made it suitable for real-time conversational AI systems. We would recommend it to teams building multilingual voice agents, contact centre AI solutions, and speech-based applications focused on regional language support.

MERaLiON has been a practical foundation for building our local language STT use cases, the model is useful, adaptable, and accessible to real-world caregivers. It has helped us move faster while keeping our stack open and extensible, which matters deeply in healthcare and eldercare. We would recommend MERaLiON to teams building applied AI products where trust, usability, and local context matter.

MERaLiON enables our solution to be localised, with a strong focus on local businesses and communities, addressing challenges faced by Singaporean. We are happy to work with the MERaLiON team.