Audio AI for Singapore
Southeast Asia's first multimodal AudioLLM — understanding speech, emotion, and language across Singapore's diverse accents, dialects, and cultures.
NewMERaLiON-3-10B-preview is now available.
Capabilities
Speech Recognition
Transcribe Singapore English, Singlish, Mandarin, and regional accents. Supports live recording and file upload, with a dedicated fast-encoder option for batch workloads.
Try transcriptionAudioLLM Playground
Explore spoken Q&A, speech summarisation, translation, and paralinguistic tasks. Includes 26+ curated audio examples across 6 task categories.
Open playgroundEmotion & Gender Analysis
Segment-level emotion and speaker gender detection from audio, powered by MERaLiON's dedicated speech emotion recognition model.
Analyse audioAPI Console
Register for API access, retrieve your key, and track usage against your tier quota. OpenAI-compatible endpoint — drop in with any existing SDK.
Get API access