Developer Access

API Console

MERaLiON API

Audio-native language model API — OpenAI-compatible, built for Southeast Asian languages.


Limitation of Liability: To the fullest extent permitted by law, A*STAR I2R or A*STAR shall not be liable for any damages arising out of or in connection with the Model, including but not limited to reliance on any Outputs.

OpenAPI Docs ↗Admin ↗
Try It Live

Make a real API request with your key. The generated curl command updates as you type.

Transcription endpoint — returns the transcript

Click to select an audio file (WAV, MP3, M4A…)
curl · live preview
curl -X POST https://api.meralion.ai/v1/audio/transcriptions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "audio_url": "data:audio/wav;base64,<BASE64_AUDIO>"
  }'

Base URL

Productionhttps://api.meralion.ai

For chat / Q&A the API is OpenAI-compatible — point any OpenAI SDK's base_url to https://api.meralion.ai/v1. For speech-to-text, call /audio/transcriptions directly (see example) — it is the correct endpoint for transcription.

Transcription Example

Speech-to-text uses the dedicated /audio/transcriptions endpoint. Use /chat/completions only for conversational Q&A about audio, not bulk transcription.

curl
# Speech-to-text: use the dedicated /audio/transcriptions endpoint.
# (Do NOT use /chat/completions for transcription.)
curl -X POST https://api.meralion.ai/v1/audio/transcriptions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "audio_url": "data:audio/wav;base64,<BASE64_AUDIO>"
  }'
Python
import base64, pathlib, requests

audio_b64 = base64.b64encode(pathlib.Path("speech.wav").read_bytes()).decode()

# Speech-to-text via the dedicated ASR endpoint
resp = requests.post(
    "https://api.meralion.ai/audio/transcription",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={
        "audio_url": f"data:audio/wav;base64,{audio_b64}"
    },
)
message = resp.json()["choices"][0]["message"]
print(message["content"])             # full transcript
JavaScript
import fs from "node:fs";

const base64 = fs.readFileSync("speech.wav").toString("base64");

// Speech-to-text via the dedicated ASR endpoint.
const resp = await fetch("https://api.meralion.ai/audio/transcription", {
  method: "POST",
  headers: {
    "Authorization": "Bearer YOUR_API_KEY",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    audio_url: `data:audio/wav;base64,${base64}`
  }),
});
const { message } = (await resp.json()).choices[0];
console.log(message.content);    // full transcript

Authentication

Recommended
Header
Authorization: Bearer YOUR_API_KEY
Header
X-API-Key: YOUR_API_KEY
Query param
?api_key=YOUR_API_KEY

Key Endpoints

POST/v1/audio/transcriptionsSpeech-to-text (ASR)
POST/v1/chat/completionsMultimodal chat / Q&A (not for bulk transcription)
POST/keys/registerRegister for an API key
GET/v1/modelsList available models
GET/keys/usageCheck your usage & limits
GET/keys/tiersView available tiers

Full OpenAPI spec: http://meralion.org:8010/docs ↗